An Introduction to Regional Economics
Edgar M. Hoover and Frank Giarratani
The Location of People


The importance of manpower supply as a location factor is suggested by the sheer magnitude of labor costs as an element in the total outlays of productive enterprises. In the United States, wage and salary payments, and supplements thereto, account for about three-fourths of the national income, and this does not include the earnings of self-employed people and business proprietors (for example, farmers, store owners, and free-lance professionals), which mainly represent a return to their labor. Accordingly, we should expect to find many kinds of activities locationally sensitive to the differentials in the availability, price, and quality of labor.

Labor’s role as a purchased input, however, is only one aspect of the locational interdependence of people and their economic activities. People in their role as consumers of the final output of goods and services affect the locational choices of market-oriented activities. They play still another role as users of residential land; and in urban areas, residence is by far the largest land use. Finally, and most important, the purpose of the whole economic system is to provide a livelihood for people. Regional economics is vitally concerned with regional income differences, the opportunities found in different types of communities, and regional population growth and migration.

The present chapter is devoted to integrating and exploring these various aspects of "the location of people." We begin by considering the locational differences in the rewards or "price" of labor.


Comparing wages or incomes among different areas is not as straightforward a matter as it might appear, even when appropriate data are at hand. It is a question of what comparison is relevant to the question we have in mind. For example, if we want to gauge the relative opulence of two communities (perhaps as an indication of how rich a market each would provide for consumer goods and services), then average personal income in dollars per family or per person would be appropriate to compare. But an individual looking for an area where his work will be well rewarded would do better to compare real earnings in his or her occupational category; that is, monetary earnings deflated by a cost-of-living index. Finally, an employer looking for a good labor supply location would be most interested in comparisons of labor cost, based on monetary wage-and-salary rates adjusted for labor productivity and fringe benefits. The relevance of such different measures should be kept in mind as we look at the data.

10.2.1 Differentials in Pay Levels

Rates of pay in any specific occupation can differ widely from one place to another and even within the same labor market area. For example, in 1978 the union hourly wage scale for laborers and helpers in the building trades averaged $8.54 for 65 cities surveyed by the U.S. Bureau of Labor Statistics, with the rate in individual cities ranging from a high of $10.54 in Cleveland, Ohio, to a low of $5.21 in Huntsville, Alabama.1

The differentials among regions are not entirely erratic. Some evidence of an underlying pattern appears in Table 10-1, in which the labor markets are classified by broad region and by size. In each of the three broad occupational categories, the South shows up as the region with the lowest pay levels. The West and North Central regions pay generally higher rates. In addition, with the exception of two occupational categories in the North Central region, there is a tendency for rates to be higher in the larger metropolitan areas. Finally, it can be noted that the interregional disparities are wider for unskilled workers than for the other groups. This feature of the pattern will be explained later.

Although the figures cited are based on careful comparisons of the standard earnings rate in (as nearly as possible) identical jobs, they do not give a complete picture of relative advantages for either the employee or the employer. No account is taken of the increasingly important fringe benefits (vacations, overtime pay, sick leave, pensions, and so on) or of differences in the cost of living. Nor do these comparisons give us any indication of differentials in the productivity of workers, which also play a part in determining the employer’s labor cost per unit of output.

10.2.2 Income Differentials

Income differentials also show a discernible pattern according to region and size of urban place. But the difference in per capita or per family incomes between two areas is, of course, determined not only by relative earnings levels in specific occupations but also by differences in the occupational and industry mix of the areas, the degree of labor force participation, and unemployment rates. For example, regional per capita personal income in 1981 varied as shown in Table 10-2.

The relation of income level to type of urban or rural place of residence is shown in Table 10-3. We observe there that in 1980, incomes were higher in metropolitan areas than in nonmetropolitan areas; higher in larger metropolitan areas than in smaller metropolitan areas; higher outside central cities than inside of central cities; and higher in nonfarm rural areas than on farms.

10.2.3 Differentials in Living Costs and Real Income

From the standpoint of the worker, the possible advantage of working in a high-wage or high-income area depends partly on how expensive it is to live there. Most of us are aware that there are considerable differences in the cost of living in different parts of the country and different sizes of community.

Although it is impossible to measure relative living costs comprehensively so as to take into account all the needs and preferences of an individual, a useful indication is provided by surveys of the comparative cost, in different locations, of securing a specific "standard family budget" of goods and services. Table 10-4 summarizes the findings of a survey of this type. A fairly distinct pattern of differentials appears. With few exceptions, living costs are higher in metropolitan areas than in nonmetropolitan areas for major kinds of expenditure. When one looks at total family consumption, living costs in the South are clearly lower than elsewhere in both metropolitan and nonmetropolitan areas. This is attributable to the comparatively low cost of housing and food in the South. Also, examination of the indices for individual SMSAs reported in the survey suggests that high housing costs are associated with large city size, rapid recent growth, and rigorous climate.

A study of SMSA characteristics associated with living cost for low-, moderate-, and high-income families in 38 SMSAs, using regression analysis, found that 64 percent of the total variance in living costs for moderate- and high-income families could be explained in terms of three significant variables: population, location in the Southeast or elsewhere, and the degree to which spatial expansion of the urban area was subject to "topological and physical constraints" (for example, water or mountain barriers) on its periphery. This last factor could be expected to influence travel distances and costs and also land cost. Climate did not show up as a significantly correlated characteristic, nor did size of place in the case of low-income families.2

A crude picture of differentials in real incomes among metropolitan areas can be obtained by dividing the per capita personal income for each area by the index of consumer budget costs for the same area.3 For the sample of metropolitan areas on which Table 10-4 is based, the real income index so obtained is rather well correlated with per capita personal money income, while per capita personal money income is somewhat less strongly correlated with the index of consumer budget costs.

Thus living costs tend to be high where money incomes are high (not surprisingly, in view of the important impact of service costs and other local labor costs on the consumer budget). But the interarea differentials seem to be wider for money incomes than for budget costs; so real income thus estimated is a little higher in places where money income is high. By the same token, interarea differentials in real income are much smaller than those in money income.

There is evidence that similar relationships prevail also when we compare different size classes of places,4 though it is impossible to compare adequately the psychic satisfactions and costs that come from living in large cities as against smaller places.


In trying to understand the causes and effects of wage and income differentials, it is useful to consider separately the supply and demand sides of the local labor market. The aggregate supply of labor in a community may be quite inelastic in the short run, since it can change only through migration or changes in labor force participation. For a single activity or occupation within the labor market area, the supply is more elastic because it can be affected by workers changing their activities or occupations as well as by migration and changes in the labor force. The labor supply as seen by an individual employer is still more elastic, and for small employers in a large labor market almost perfectly so.

10.3.1 Work Location Preferences and Labor Mobility

There are many reasons for preferring a job in one area to the same kind of job in another area, and the decision to move can be very complex. A systematic evaluation of costs and benefits is in order, much as a decision-maker in business evaluates the pros and cons of an investment. Thus the potential migrant would want to recognize what he or she would be giving up (the opportunity costs of the move) and make a good guess about what would lie in store in the new location.

Neither of these tasks is particularly easy, and a number of considerations would have to be recognized. First, it is not adequate to compare only basic wages; all fringe benefits must be considered as well. Also, for an increasingly large share of the work force, the employment prospects of the spouse may be as important as that of the "primary" wage earner.5 Second, a community with cheaper living costs might be preferred in the absence of any pay differential. Third, various aspects of the quality of the job, such as security and prospects of advancement, may be considered. Expected growth in earnings some years down the road may be an important factor in the decision to act now.6 Finally, other aspects of the desirability of the community as a place to live can include, for example, climate, cultural and social opportunities, and access to other places that one might like to visit.7 Differences among places on any of these accounts can be compared to money income differentials: How much additional compensation is required to give up immediate access to cultural events, or how much less would one be willing to accept in terms of earnings for a climate that suits one’s tastes?

Spatial mobility refers to people’s propensity to change locations in response to some measurable set of incentives, identified in practice as "real income" or simply as money wage rates deflated by a cost-of-living index. If mobility in this sense were perfect and real wages thus equal everywhere, there would be differentials in money wages paralleling the differentials in living costs. A labor market where the living costs were 10 percent above average would pay wages 10 percent above average, but real wages there would be the same as anywhere else.

The term equalizing differentials has been applied to this kind of money wage or income differential.8 The pattern of actual money differentials is, then, made up of two components: (1) equalizing differentials, which would exist even in the absence of any impediment to labor mobility, and (2) real differentials, representing differences in real income and thus presumably caused by impediments to mobility.

The significance of this distinction is that workers can logically be expected to choose locations and to move in response to real differentials; whereas employers looking for cheap labor will be more interested in the total money wage differential, which combines both real and equalizing differentials.

These concepts of real and money wages, cost of living, equalizing and real differentials, and mobility help us to understand some basic motivations of the locational choices of employees and employers; but unfortunately they are not very sharp tools. We have already noted the impossibility of including in indices of income and living costs all the considerations affecting the desirability of a place to live. For example, such indices take no account of the attractions of a mild and sunny climate (except as reflected in housing costs), the dirt and discomforts of life in a large industrial city, the social pressures and cultural voids of a small town, or the advantage to a research worker of being stationed where the action is in his or her field. As a result of our inability to measure real income fully, we are also unable to measure mobility in a completely unambiguous way. If a family, for example, likes the physical and social climate of its surroundings and refrains from moving to another area where the pay is higher both in money terms and as deflated by a conventional family budget cost index, should we ascribe its failure to move to a lack of mobility?

A further difficulty with the simple concept of real and equalizing differentials is the implication that migration is not merely motivated by real-income differentials but tends to eliminate them. Under certain conditions it is possible for migration, even when so motivated, to leave the differential unchanged or even to widen it. We need to look further into both the causes and the consequences of migration.

10.3.2 Who Migrates: Why, When, and Where?

Migration is influenced by three conditions: the characteristics of both the origin and destination areas, the difficulties of the journey itself, and the characteristics of the migrant.

Reasons for Moving. It is a drastic oversimplification to explain migration simply on the basis of response to differentials in wage rates, income, or employment opportunity. The U.S. Bureau of the Census bases its tabulations of migration on changes in residence. Since some of these are local (moves within the same county) and others involve substantial distance (intercounty moves), one would expect reasons for moving to differ widely.

Those persons who move only within the same county are predominantly influenced by housing considerations. Since all of a county is generally regarded as being included within a single labor market or commuting range, job changes are related only to a minor extent with intracounty moves. Most such movers are not changing jobs.

For those who move to a different county the picture is quite different, with employment changes (including entry to or exit from military service) emerging as the major reasons for migrating. This reflects the fact that an intercounty migration generally involves shifting to a different labor market beyond the commuting range for the former job. A change of residence is involved but is not the primary motivation.

Characteristics of Origin and Destination Areas. The characteristics most obviously affecting attractiveness to the individual migrant have already been suggested. In addition, we should expect that a larger place would have more migrants arriving and departing than a smaller place, more or less in proportion to size. But size itself can significantly affect the appeal of a place for the individual. Historically, migrants have responded to the greater variety of job opportunities in larger urban places and their suburbs by migrating to them in numbers somewhat more than proportional to their size.

Rather than thinking of migration as being motivated simply by net advantages of some places over others, it is useful to separate the pull of attractive characteristics from the push of unattractive ones. Clearly, many people migrate because they do not like it where they are, or perhaps are even being forced out by economic, political, or social pressures. The basic decision is to get out, and the choice of a particular place to migrate to is a secondary and subsequent decision, involving a somewhat different set of considerations. On the other hand, some areas can be so generally attractive as to pull migrants from a wide variety of other locations, including many who were reasonably well satisfied where they were.9

Recent studies on migration using detailed data on flows in both directions (rather than just net flows) have considerably revised earlier notions of push and pull. Rather surprisingly, it appears that in most cases the so-called push factor explaining out-migration from an area is not primarily the economic characteristics of the area (such as low wages or high unemployment) but the demographic characteristics of the population of the area. Areas with a high proportion of well-educated young adults have high rates of out-migration, regardless of local economic opportunity. The pull factor (that is, the migrant’s choice of where to go) is, however, primarily a matter of the economic characteristics of areas. Migration is consistently heavier into prosperous areas. Accordingly, the observed net migration losses of depressed areas generally reflect low in-migration but not high out-migration, and the net migration gains of prosperous areas reflect high in-migration rather than low out-migration.10

Difficulties of the Journey. Within a country, distance is perhaps the most obviously significant characteristic of the migration journey, and virtually all analyses of migration flows have evaluated the extent to which migration streams attenuate with longer distance. Thus a simple "gravity model" of migration posits that the annual net migration from A to B will be proportional to the populations of A and B and to the size of some differential (say, in wage rates) between A and B, and inversely proportional to the square of the distance from A to B, as follows:

(where the Ps represent populations, the Ws wage rates, D the distance, M the number of migrants per unit of time, and g a constant with a value depending on what units are used for the variables).11 This basic migration flow model has been statistically tested and modified in many ways in the attempt to make it more realistic. It has already been suggested that in many circumstances at least, there is a "scale effect" upon migration, which can be incorporated in the model by giving the populations an exponent greater than 1. Any relevant factor of differential advantage that can be quantified (for example, unemployment rates, mean summer or winter temperature, percentage of sunny days, average education or income level of the population, percentage of housing in good condition, crime rates, insurance rates, or air pollution) can be introduced, with whatever relative weighting the user of the model deems appropriate.

The distance factor in the model likewise can be assigned a different exponent to fit the circumstances (there is nothing special about the square of the distance, except for the law of physical gravitation) and elaborated in various ways. Actually, distance per se is at best only loosely related to the difficulties attending migration. The factors involved are actual moving costs (which can sometimes be the least important obstacle), uncertainty, risk and investment of time involved (including that associated with acquiring information), restrictions on migration per Se, and what is sometimes called social distance—suggesting the degree of difficulty the migrant may have in making adequate social adjustment after he or she arrives.12 As an example of this last factor, a model designed by W. H. Somermeijer to explain Dutch internal migration flows included a term that measured the difference in the Catholic-Protestant ratio between the two areas. Introduction of this term substantially improved the model’s explanatory power, suggesting that members of each religious persuasion tend to move mainly to areas where their coreligionists predominate. 13

Social distance depends partly, of course, on the individual migrant. But wide social distances between communities and regions (that is, great heterogeneity) tend to restrict migration flows to those individuals who can most easily make the required adjustment. With the improvement in communications and travel, social distance in space tends to lessen; but it is still a factor, for example, between French and English Canada, between the North and the Deep South in the United States, or between farm and city.14

Another feature of migration paths is that they seem to be subject to economies of volume of traffic along any one route. Well-beaten paths become increasingly easy and popular for successive migrants. This is so for a variety of reasons. Sometimes (as in the case of earlier transoceanic and more recent Puerto Rico-to-mainland migrant travel) transport agencies have given special rates or in other ways have favored the increase of migrant travel on the most frequented routes. Perhaps more generally applicable and more important is the fact that migrants try to minimize uncertainties and risks by choosing places about which they have at least a little information and where they will find relatives, friends, or others from their home areas who will help them to gain a foothold.15 This tendency is particularly important, of course, when the social distance is large. It goes far to explain the heavy concentration of late-nineteenth-century European migrants to the United States in a few large cities, and the even more remarkable concentration of particular ethnic and sub-ethnic groups in certain cities and neighborhoods, which even now retain unto the third generation some of their special character. The most recent ethnically distinctive waves of migration to American cities—those of blacks, Puerto Ricans, and Chicanos—have in the same fashion followed a few well-beaten paths. For example, it has been established that Southern black migrants to Chicago came mainly from certain sections of the South, while those going to Washington, D.C., or Baltimore originated in other Southern areas, and those going to the West Coast in still other areas. Recent migrants from the Caribbean (other than those from Puerto Rico) remain highly concentrated in the Miami area.

An analysis of labor mobility in the United States between 1957 and 1960, based on Social Security records, gives striking evidence of the beaten-path effect where migration of blacks is concerned. For white workers, both male and female, migration flows were significantly related to earnings differentials (positively) and to distance (negatively) as we might expect. For black men, however (and to a lesser extent for black women), earnings differentials and distance appeared to be less important determinants than either the number of blacks or the proportion of blacks in the work force of the destination area. In other words, blacks (or at any rate, black males) were selectively attracted to labor markets in which there already was a high proportion of blacks.16

The tendency of migration to channelize in well-beaten paths provides part of the explanation for another characteristic of migration streams already mentioned; namely, that places with high rates of inward migration tend to have high rates of outward migration as well. This so-called counterstream effect was noted as long ago as the 1880s in E. G. Ravenstein’s pioneer statement of migration principles, and has been amply verified since.17

The point here is that a well-beaten path eases travel in both directions. Migrants generally come to a place with incomplete knowledge, and many of the disappointed ones simply retrace their steps. Also, a city that offers especially favorable adjustment opportunities for migrants is likely to serve as a "port of entry" for migrants coming into that region or country. New York has historically been the entry place for transatlantic immigrants, and Chicago has functioned as an important first destination for Mexicans migrating to the American Midwest. In such cases, many of the migrants move out again (either to other places in the region or perhaps back to their place of origin), so the out-migration rate of such an entry point or staging area is likely to be high. More generally, it is plausible to assume that people who have just migrated are especially mobile by circumstances or taste and hence more likely than others to swell the outward flow.

Migration flows over long distances within the same country have historically involved a considerable amount of what is sometimes called chain migration. Most of the migrants move relatively short distances, but the moves are predominantly in one direction, forming a stream. As people move from B to A, they are replaced in B by migrants from C, who in turn are replaced by migrants from D. It has been established that most of the massive redistribution of population in England during the Industrial Revolution was carried out in such fashion by short-distance moves cumulating toward the new industrial towns.18

Characteristics of the Migrant. Migration is basically selective. Some people are far more prone to migrate than are others. This is often expressed as a difference in the mobility of different groups, but we really cannot explain all of the difference in that way, since the incentives to migrate are not the same. For example, a young scientist fresh out of school is confronted with a quite different set of pushes and pulls than an aging farmer, an established business executive, a manual laborer, or a wealthy widow.

The most conspicuous differences in migration rates are those experienced by the individual in passing through successive stages of a lifetime. These changes are (as shown in Figure 10-1) rather similar for the two sexes; for simplicity’s sake we describe them here in male terms.

A very young child is relatively portable. After he enters school and has older, more settled parents and probably more brothers or sisters, his probability of moving declines. The rate rises suddenly when he is ready to look for a job or choose a college, and it remains high until after he is married and has children of his own. As his stake in his job and community grows, and as his and his family’s other local ties develop, he becomes less and less likely to move. His mobility recovers somewhat at the stage when all of his children are on their own, and again at the customary mid-sixties retirement age. After about age seventy, migration rates tend to rise a little, presumably reflecting adjustments to death of the spouse or to growing incapacity.

These characteristic life-cycle variations are manifest in Figure 10-1, showing United States migration rates by age and sex. The pattern is blurred in the aggregate, of course, by the fact that not all people enter the labor force, marry, or retire at the same ages. But age remains the characteristic most distinctly associated with migration-rate differentials. Most of the migration that occurs (except in massive displacements of populations by military or political force or by natural disaster) is done by people in young adult age groups.

Certain other individual characteristics also substantially affect migration rates. The most important are marital status, parenthood, and level of education. Table 10-5 shows migration rates over a five-year interval for men in three mature age groups (that is, after their schooling was probably completed), according to amount of schooling. We see that higher education is associated with higher migration rates. Occupational status also has a bearing on migration rates. For a given age category, individuals with higher occupational status have greater mobility.19

In addition to such regularly recorded characteristics as we have considered, individuals have many other personal characteristics that influence their propensity to migrate. Some people are simply more footloose, more adventurous, more easily dissatisfied, or more ambitious than others, or in countless other ways more mobile.

These wide differences in migration rates among different kinds of people mean, of course, that those who migrate are almost never a representative cross section of the population of either their area of origin or their area of destination. A migration stream substantially alters the make-up of the population and the labor force in both areas.

The selectivity of migration is greatest when the journey is difficult, when the areas of origin and destination are in sharp contrast, and when the population itself is highly diverse in such characteristics as education, income level, occupational experience, and ethnic or racial background. Migrants generally seem to be somewhat above the average of the origin area in terms of energy, ability, and training; this suggests that the direct effect on the remaining population is to lower its average "quality." One of the oldest clichés regarding emigration, in fact, is that it tends to drain the best people out of an area, thus damaging the prospects for industrialization or other economic development.

An intensive study of the occupational status of American men aged twenty to sixty-four came to these conclusions:

Migration has become increasingly selective of high potential achievers in recent decades.
…The careers of migrants are in almost all comparisons, clearly superior to those of nonmigrants. . . . Whether migration between regions or between communities is examined; whether migrants are compared to nonmigrants within ethnic-nativity groupings or without employing these controls; whether education and first job are held constant; and whether migrants are compared to natives in their place of origin or their place of destination— migrants tend to attain higher occupational levels and to experience more upward mobility than nonmigrants, with Only a few exceptions.
…Migrants from urban places, though not those from rural areas, enjoy higher status than the natives in the community to which they have come, regardless of its size.20

These findings go well beyond the traditional folklore about selective migration in that they suggest (1) that migrants (except from rural areas) tend to be superior to the population of the destination area as well as of the origin area, and (2) that selectivity may be increasing, contrary to the expectation that more general literacy and other trends enhance the mobility of an increasing proportion of the population.

Another study, relating to migrants to the industrial metropolis of Monterrey, Mexico, between 1940 and 196021 gives a rather different picture. Migration appears to have become much less selective with respect to populations of origin in that interval. Early in the period, only a few of the more highly educated ventured the move to the city; later, mobility increased, so that villagers of all educational and income levels became more nearly representative of the populations of its areas of origin, while at the same time it was becoming increasingly different from (educationally inferior to) the population of the urban destination area.22

Migration selectivity, then, can depend to a large extent on the state of development of the regions involved and can change in character fairly rapidly. It has already been suggested that a broadening of education and economic opportunity in less-developed regions can reduce selectivity. Finally, Everett Lee has proposed a plausible but not easily verifiable hypothesis: that migration motivated by pull tends to be positively selective (i.e., the more productive people are the ones who go), whereas migration motivated by push tends to be negatively selective.

Factors at origin operate most stringently against persons who in some way have failed economically or socially. Though there are conditions in many places which push out the unorthodox and the highly creative, it is more likely to be the uneducated or the disturbed who are forced to migrate.23

Inward, Outward, and Net Migration: Three Hypotheses. The typical age selectivity of migration is sometimes an important factor accounting for the observed tendency that places with high in-migration rates have high out-migration rates as well. The most mobile age groups (and perhaps also the individuals most mobile by temperament or other characteristics) are present in abnormally high proportion in a fast-growing area with high recent and current in-migration; and such local demographic characteristics play a large part in determining how many people leave an area.24

We have learned that the relationships among inward, outward, and net migration are more complex than one might suspect. Let us review them in terms of three hypotheses or "laws of migration," using Figure 10-2 as a graphic aid.

The naive or common-sense" expectation regarding migration into and out of an area is that if the area is attractive as a place to work and live, there will be a net inward flow reflecting large inward and small outward migration; while if the area is unattractive, there will be a net outflow reflecting large outward and small inward migration. This hypothesis is represented diagrammatically in the first panel, (a), of Figure 10-2, where the dots could represent different labor market areas or the same area at different times. In-migration and out-migration rates are measured on the horizontal and vertical axes respectively. The 45-degree line represents zero net migration. The various areas show a pattern of negative correlation of out-migration, with both inward and net migration: "Attractive" areas are those in the lower right part of the scatter, and "unattractive" areas are those above and to the left of the diagonal.

The second panel, (b), depicts the contrasting relationship, previously suggested on the basis of migration selectivity and other factors, which we might designate as the Lowry hypothesis. Here the rate of out-migration is positively correlated with both inward and net migration.

Still another view is that shown in the third panel, (c), which we may call the Beale hypothesis.25 Using 1955-1960 data for 509 State Economic Areas demarcated by the Census Bureau,26 Beale discovered a relationship schematically resembling that of Figure 10-2 (c). He found that high gross out-migration can be associated with either high net in-migration or high net out-migration. The net rate is mainly determined by out-migration when the net is negative, and by in-migration when the net is positive. We may interpret this as meaning that the "Lowry effect" dominates in relatively prosperous and growing areas; whereas in areas that are seriously depressed the predominant effect is the "common-sense" one: Poor prospects both discourage inflow and encourage outflow, and economic factors exert both pull and push.

The significance of this issue is far from trivial, as Beale points out and as we shall more fully appreciate in the context of Chapters 11 and 12. If migration out of seriously depressed or backward areas is assumed to be affected only by demographic characteristics of the population and not by the level of unemployment or income, then measures to stimulate activity in such areas would not reduce out-migration (in fact, according to the Lowry hypothesis they would eventually increase it). The Beale findings suggest, however, that such stimulus may, in some cases at least, retard out-migration. In choosing among policy decisions regarding aid to depressed or backward areas, it is of some importance to try to gauge this possible impact.

Changes in Migration Rates. There are a number of reasons why one might expect migration rates to increase over time. Lee has suggested that increasing diversity of the opportunities afforded by different areas, increasingly diverse specialization of people’s capabilities and preferences, the beaten-path effect, and the increasingly wide knowledge and experience of other locations that is brought about by education, better communication, more income, and increased leisure to travel would each enhance the mobility of the population.

Changes in the socioeconomic and demographic characteristics of households have been working in the opposite direction, however. For short-distance (intracounty) moves, decreases in average size of household and increases in home ownership are associated with decreases in mobility. With respect to long-distance moves, the rising incidence of two-wage-earner households also discourages mobility.

Recent tabulations of migration data gathered on an annual basis from sample surveys since 1948 show that the net effect of these countervailing forces has been a consistent decline in the overall migration rate. For example, the average annual rate for the twelve-month period March to March has fallen from 20.6 in 1961-1962 to 18.7 in 1970-1971 and to 17.2 in 1980-1981. Thus in the 1960-1961 period, roughly 21 percent of the population changed residences in the United States, whereas that number had fallen to about 17 percent in the 1980-1981 period.27

Much of this trend can be attributed to decreases in the frequency of intracounty moves. However, there are also indications that regions have become more alike, so that in some respects the incentives for long-distance moves have probably lessened. Regional incomes in any case have tended to converge toward the national average. Thus in addition to the socioeconomic and demographic factors mentioned above, reduced migration incentives could also have offset increases in personal mobility.

Migration rates show marked seasonal and cyclical variations. In general, prosperity favors migration because opportunities are more plentiful, risks of unemployment at the new location are less, and migrants themselves are in a better financial position. In periods of economic recession or depression, people tend to look for the place with the best economic security. This may mean staying where they are or (in the case of fairly recent migrants) returning to their last place of residence. Thus the long-term migration stream toward places with better long-term prospects is temporarily interrupted or even reversed by severe recessions. For example, in the worst years of the Great Depression of the 1930s, the longstanding net flows of migrants from farm to nonfarm areas and from foreign countries to the United States were both temporarily reversed.

The Effectiveness of Migration. When people migrate, they are seeking to better their prospects. How well does migration accomplish that purpose, and how does it affect people other than those who migrate?

The answer obviously depends in part on how accurately people size up the prospects when they decide to move (or not to). The better informed they are, the greater the probability that migration will justify itself and will contribute to a more efficient allocation of human resources in the economy as a whole. Thus public policy with regard to migration should, first, help potential migrants to get the information they need for rational choice. Beyond this rather obvious point, the question of the effectiveness of migration can be examined on three different levels.

  1. The "efficiency of migration" between any two areas is sometimes defined as the ratio of the net flow to the total gross flow in both directions. In other words, if all migrants go in the same direction, the efficiency is 100 percent, in the sense that there is no cross-hauling of migrants: The net flow equals the gross flow. At the other extreme, if the flows in the two directions just balance, so that there is no net movement at all, the efficiency is said to be zero.

    This measure may be useful in suggesting the degree to which a net migration figure can be misleading as an indicator of the amount of movement; but it has little to do with efficiency in any meaningful sense. People are not simply interchangeable units of manpower, as the efficiency ratio implies. On the contrary, those moving in one direction may be presumed to differ qualitatively from those moving in the other; with each stream believing, and perhaps correctly, that it is going in the right direction. Accordingly, a situation in which two opposing flows largely cancel out in terms of numbers of people is not necessarily indicative of any "lost motion" or waste in terms of either the welfare of the individual migrants or the socially desirable spatial allocation of manpower resources.

  2. A different and somewhat more sophisticated question about migration is to ask whether it seems to be going to the right places so far as the economic benefit to the migrant is concerned. If we find people moving predominantly from places of lower incomes to places of higher incomes, or from areas of heavy unemployment to labor shortage areas, we surmise that they know what they are doing. If they go the other way, or simply in all directions without any apparent regard to income or employment differentials or any other obvious index of advantage, we have to surmise either that the migrants are ignorant about the alternatives or that we are ignorant about their real motivations.

    Actually, most migration flows do fit a "rational" pattern in relation to observable differences in earning levels, unemployment rates, and such other easily identified variables as climate and the level of public assistance benefits. Even the migration of poor blacks from Southern farms to Northern city ghettos with high unemployment and dismal living conditions can make sense in terms of improvement in the migrants’ incomes, at any rate if we ignore possible adverse effects on the social adjustment of the individuals and communities involved.
    In relating migration to indices of community economic welfare (for example, wage or income levels) we cannot reasonably assume that the migrant immediately fits into the pattern of the area and receives the average pay, employment security, and other perquisites of the residents in his or her age and occupational category. Migrants are no more representative of the populations of their destination areas than they are of the populations of their areas of origin. Some of their distinctive characteristics are subject to modification, so that if they stay they will tend to become more similar to their new neighbors. This tendency toward assimilation applies to skills, consumption patterns, ratings on most kinds of "intelligence" tests, desired number of children, and social behavior.

  3. Finally, we can judge migration on the basis of how much it contributes to aggregate output or, more broadly, to general social welfare. This is the appropriate level of judgment for public authorities and public-spirited citizens to try to use. It calls for assessing the effects of migration not just on the migrants but on the communities they leave and enter.

    This is by no means a simple criterion to apply. On the face of it, the transfer of manpower to places where its productivity is higher seems likely to raise national per capita output and increase "aggregate welfare," insofar as that term has any meaning. And differentials in real-earnings rates reflect, roughly at least, differentials in the marginal productivity of labor. But migration (particularly highly selective migration) can have important side effects (externalities) on the areas involved, in terms of the costs of public services, the prospects for future economic development, and the quality of life. The discussion of these problems will be taken up in later chapters where we come to grips with the processes of regional growth and change.


Our discussion of mobility and migration has shed some light on what determines the supply of labor at different places. To the extent that people move to places offering more jobs or higher earnings, the labor supply adjusts to the spatial pattern of labor demand. In areas where demand has grown relative to supply and there are hindrances to inward migration, a tight labor market is manifest by low unemployment rates and relatively high earnings; in places where labor demand has declined or has failed to keep pace with the growth in the labor force (resulting in part from natural increase of the population) and outward migration is not easy, we find a labor surplus manifest in high unemployment rates and relatively low earnings rates. This is, of course, a somewhat simplified picture; many areas, at any given time, do not fit wholly into either of these two contrasting categories.

On the demand side, we envisage employers of labor as being concerned about its costs and seeking to make profitable adjustments to such labor cost differentials as they are aware of.

What, then, will employers do if labor is expensive (relative to its productivity) in a specific location? They have three possible ways of economizing on this expensive labor: changing their production techniques so as to substitute other inputs (for example, labor-saving machinery) for manpower; going out of business; or moving to a different location where labor is cheaper. Where the last two choices are seriously considered, we can say that the activity in question is locationally sensitive to labor supply, or to some degree is labor-oriented.

The degree of labor orientation varies widely among activities. In one extreme case (activities tightly tied to a locality by market orientation, orientation to inputs other than labor, or some other compulsion), labor costs may have no significant locational effect at all—the employer’s demand for labor at that location is highly inelastic. Retail trade and local services illustrate this category of locally bound industries essentially unaffected by labor cost differentials. If, for example, drugstore clerks were paid twice as much in Milwaukee as in Akron, this would not induce Milwaukee drugstore proprietors to relocate to Akron. Their market is entirely local. They are not in competition with Akron in any sense and only need to assure themselves that they are not paying their clerks more bounteously than their Milwaukee competitors. There would, however, be a stronger incentive in Milwaukee than in Akron to skimp on labor and substitute other inputs if possible. In the case assumed, we might expect Milwaukee drugstores to be quicker to install such things as vending machines, change-making cash registers, and display layouts facilitating self-service.

At the other extreme, we have strongly labor-oriented activities, those whose demand for labor at any particular location is highly elastic—unless labor is cheap, they will close down or go elsewhere. Normally, these are activities that are rather footloose with respect to location factors other than labor supply. A change of location makes relatively little difference in their costs of transfer, level of sales, or outlays for other local inputs per unit of sales, while their labor costs do vary markedly from one location or region to another. Historically, the manufacture of textiles and standard clothing has been strongly oriented to low-wage labor, while labor-intensive activities requiring scarce special skills have been strongly oriented to the few places where such skills are available.

Until rather recently, most activities oriented to cheap labor as such mainly employed unskilled or semiskilled blue-collar workers; but nowadays there are numerous instances of firms moving to places where there is cheap white-collar clerical labor. This is partly the result of the rapidly increasing proportion of white-collar to total employment; but the locational effect also reflects the increased availability of qualified clerical help in small communities.


Having briefly considered the location of manpower from both the supply side and the demand side, we now have some insight into the rather complex set of interrelations shown schematically in Figure 10-3. This diagram may be useful in reminding us of the interdependencies and tracing the repercussions of different kinds of change. For example, the mechanism implied by the concept of equalizing differentials in wages (that is, the tendency of migration to eliminate real-income differentials) involves the simple feedback sequence of effects shown by the solid lines in the diagram below.

A more complex and realistic model of the equalization process, taking into account the fact that the price of labor affects living costs through the price of locally produced goods and services, would include in addition the effects shown by dashed lines in the same diagram. In either model, the equalization effect is finished when there are no longer any real-income differentials (that is, equilibrium has been reached, as far as the employee is concerned). The reader may find it useful to trace out in a similar fashion, using Figure 10-3 as a guide, what happens as labor-oriented employers shift their hiring to areas of low labor cost.

10.5.1 Where Are Labor Costs Low?

What are the types of location to which a labor-oriented activity is attracted? There is no single, simple answer to this question, mainly because different activities and individual firms are seeking different kinds of labor cost economy. In some jobs, one worker can perform as well as another. For such a job, labor is a rather homogeneous input, and the wage rate is a good measure of labor cost. In other occupations, skills and aptitudes vary widely, and a poor worker is not a bargain at any wage. In some activities, the nature of the product, the type of work, and the volume of output are highly changeable, and the employer wants to be able to arrange with a minimum of difficulty for changes in job specifications, short layoffs, overtime, and other changes. In such an activity, good labor supply locations may be those where the local pool of labor is large, where the average age and seniority of workers is low, or where union bargaining has not built up a rigid structure of work rules.

Each activity, then, will have its own preferences among labor supply locations, determined by the relative emphasis it puts on low wages, skill or trainability, and flexibility.

Low wages are most often found in relatively backward areas where the demand for labor has not kept up with the natural increase of the labor force. Obviously, these are areas where manpower is impounded, as it were, by its imperfect outward mobility. Less obviously, such areas tend to develop certain characteristics that impede out-migration.

Many such areas specialize heavily in kinds of employment for which the demand has grown slowly or declined (for example, general farming or coal mining). This may, in fact, be the chief reason why they are areas of labor surplus. But this specialization also means that their labor force lacks experience in more dynamic industries or occupations, which is a disadvantage in seeking work elsewhere. There are attitudinal barriers, too, to giving up an occupation in which one has acquired skill and seniority in order to start near the bottom in a new trade. Derelict coal-mining villages in Appalachia are full of middle-aged and older ex-miners who are slow to consider any alternative line of work, even though it may be clear that the local mine is closed for good. Finally, the very existence of a labor surplus and low wages helps to lower the cost of such major budget items as shelter and services, and this helps to diminish the economic incentive to move out.

The advantages of experience and skill in a labor supply are most often found in areas where the educational level is high and where activities requiring some special skills have been concentrated for a long time. The supply of some types of highly paid and scarce manpower (such as scientific and other specialists, or persons of high artistic capability) is coming to be located increasingly in areas and communities of high physical and cultural amenity, since those types of people are in such demand that they can afford to be quite choosy about where they are willing to live. It is no accident or whim that has located so many advanced-technology and research-oriented activities in pleasant places near major universities.

Flexibility and diversity of labor supply are most likely to be found in a large labor market in an intensively urbanized region, although one might surmise that the rapid population growth of nonmetropolitan areas during the 1970s has increased the diversity of labor supply in many less developed places as well. In some large urban labor markets, however, the potential labor economies of size and diversity are partly offset by the greater rigidity of union work rules and bargaining practices, and the greater age and seniority of the work force that characterize an area where an activity has developed a mature concentration.

The different kinds of labor cost advantage (low wage rates, skill, and flexibility and diversity of supply) are unlikely to be found in the same places, because to some extent they reflect contrasting area characteristics. Thus low wage rates are associated with less developed areas having low living costs. Experience and diversity of supply are often associated with the opposite type of location: large urban areas in advanced industrialized regions having relatively high living costs. Any given area, then, tends to be classified according to the aspect of labor supply in which it has the greatest comparative advantage.

10.5.2 Indirect Advantages of Labor Quality

The cost savings involved in the use of highly productive (that is, skilled and/or adaptable) manpower are sometimes underestimated, because not all of them show up in labor cost per se. Recall the case of Harkinsville and Parkston discussed in Chapter 2. Harkinsville workers work faster, and get correspondingly higher hourly wages, than those in Parkston. Thus there is no difference in labor cost per hour. Nevertheless, it will be recalled, Harkinsville is a better location for the employer. For example, if the output of the firm is to be 1000 units a day at whichever location is chosen, the production worker payroll will be the same at either location; but the plant can be smaller at Harkinsville. This means a smaller investment in land and buildings; a smaller parking lot and cafeteria; fewer washrooms, drinking fountains, and other facilities; a smaller work load for payroll accounting and personnel management; and so on. Only such cost items as are geared directly to the volume of output and not the number of people working will be as large in Harkinsville as in Parkston: for example, production materials, shipping containers, motive power and fuel for processing, and loading and handling facilities.

10.5.3 Institutional Constraints on Wages and Labor Costs

To an increasing extent in most countries, the supply of labor to an individual employer at one location is affected by bargaining procedures and constraints involving other employers and other locations as well. In activities where the employing firms are few and large and where a strong labor organization includes a major fraction of the workers, key negotiations can set a quasi-national pattern of fringe benefits and work rules subject to only minor local differences. Examples include the steel and automobile industries and rail and air transportation. In still other activities, agreements cover major sections of the activity (such as the East Coast ports with respect to handling ship cargo).

Multiarea bargaining introduces a strong additional equalization element into the wage pattern. Even more generally, labor organizations with aspirations for nationwide power try to work toward elimination of regional wage differentials. Lower wages in areas of weaker organization are viewed as a threat to employment in the areas of stronger organization and higher wages, since employers are naturally tempted to move to save labor costs. And employers not contemplating such a move are, of course, in favor of higher labor costs for their competitors. Both parties, then, may well favor extension of the geographical area of wage bargaining and the enactment of federal and state minimum-wage laws, which limit differentials still further.

There are often pressures toward similarity of pay rates and fringe benefit levels among different activities and occupations within a single-labor market. This means that if there are important high-wage activities in an area, they tend to some extent to set the tone for related types of employment in the same area and to make it generally a higher-cost area than it might otherwise be.

It is easy to see how this works between occupations calling for similar qualifications, so that the various activities in the community are competing in a common pool of available workers with those qualifications. For example, steel workers are relatively highly paid; as a result, a community dominated by steel making is likely to have relatively high wage scales in construction and in other kinds of employment not too different in their requirements from the jobs of many steel mill employees. There would be no such direct effect, however, on wage rates for sales or clerical workers.

Furthermore, unions of the dominant industry in a community generally organize a number of other industries as well, the jurisdictional lines being rather loose. Thus in Pittsburgh, metal fabricating plants are mainly organized by the steelworkers’ union, while similar plants in Detroit have locals of the automobile workers’ union. Though such a union does not necessarily find it feasible or desirable to extend the wage and benefits pattern in the dominant industry to the other industries it has organized in the same community, there is certainly some pressure in that direction, and a consequent reinforcing of the tendency toward intraarea wage level conformity.28

This tendency is not area-wide, as a rule. For example, in the Pittsburgh labor market area the upward wage pressures arising from the importance of some tightly organized and high-paying industries are essentially restricted to blue-collar occupations traditionally dominated by male workers, which heavily predominate in the major industries involved. For many years, however, Pittsburgh wages in retail trade and generally for jobs traditionally held by women tended to be a little lower than those in cities of comparable size in the same part of the country. Presumably, this reflected the relatively slow growth of the area as a whole and the relative surplus of employable women arising from the predominance of male jobs in the area. It is interesting to note that in the late 1950s the differentials began to disappear, perhaps reflecting vigorous local growth in office employment and no growth in heavy industry employment.29

It has been observed that the wage spread or skill margin among different occupations within a single labor market is generally wider in less developed and slower growing regions. This has been explained in terms of the lower educational standards and the smaller proportion of semiskilled manufacturing jobs in such areas, since education and the availability of an accessible "ladder" of skill development both enhance occupational mobility and make the labor market more competitive.30

A further explanation lies in the tendency for people of higher occupational, educational, and earnings levels to be geographically more mobile. As a result, regional differentials in their earnings are narrower than is the case for lower-status people.31 In an area of labor surplus and out-migration, it is the people in the better-paid occupations who move out most readily; a relatively larger differential is required to move the unskilled. The wage spread in such a labor market is consequently wide compared to that in a more prosperous and active place.

10.5.4 Complementary Labor

Different categories of labor are to some extent jointly supplied; that is, the supply of one kind of labor in an area depends on how much of the other kind is there. A local population or potential labor force is almost always an assortment of people of different ages, sexes, and physical and mental capabilities. If most of the jobs available in the area call for superior aptitude, the area is likely to have a surplus of people with more pedestrian abilities, and these may represent a bargain in labor supply for an activity with less exacting requirements. Conversely, if all the jobs in an area involve rugged manual labor, there is likely to be a surplus of not quite so rugged individuals, who might represent a bargain labor supply for an activity not requiring physical strength.

We have to consider, then, as another kind of advantageous labor location for an activity, places where there is a heavy demand for some kind of contrasting and complementary labor. This principle assumes, of course, that mobility is quite imperfect, so that not all the different types of workers are able to seek out the locations where their own kind of work is best rewarded.

The most important basis for such restriction is inherent in family ties. In a family’s choice of location from the standpoint of income, the most important consideration is opportunity and earnings for the principal earner, usually the head of the family. The spouse, and any other full or partial dependents of working age, is then part of the potential labor supply in the area where the principal wage earner locates. Accordingly, a labor market heavily specialized in activities employing men is likely to have a plentiful and relatively cheap female labor supply, and a labor market heavily specialized in activities employing mature adults is likely to have a plentiful and relatively cheap supply of young labor of both sexes.

Historically, many industries have owed their start in certain areas to a complementary labor supply generated in such a way. A classic case is the making of shoes in colonial days in eastern Massachusetts. The coastal area north of Boston was heavily specialized in the male occupations of sailing and fishing; this created a large complementary labor surplus of wives and daughters, who were able to supplement family incomes by making shoes—first at home and later in small factories.32 At a later period, the anthracite mining area of eastern Pennsylvania attracted a substantial amount of the silk-weaving industry on a similar basis. As late as the 1920s, it was observed that in the anthracite area about 60 percent of the silk weavers were women, while in the previous center of that industry (in and around Paterson, New Jersey) 60 percent were men. The manufacture of cheap standard garments, cigars, and light electrical equipment and components also has historically been attracted to places offering plentiful and cheap complementary labor.33


Up to this point, we have been looking at the location of people and differentials in earnings and labor costs on a macrogeographic scale, comparing labor markets as units. A quite different set of considerations comes to the fore when we adopt a microgeographic focus. Since the question of people’s residential location preferences within urban areas has been dealt with in some detail in Chapters 6 and 7, we shall focus on the locational preferences of employers as influenced by labor supply.

Although labor market areas are in principle defined in terms of a feasible commuting range, people prefer short work journeys to long ones, and as shown in Chapter 6, residential location decisions reflect this and other access considerations. Thus the supply of labor to an employer is not really ubiquitous throughout an urban area. Differential advantages of labor supply within a local labor market are particularly significant when the employer wants a special type of labor or a large supply of job candidates, and in large labor market areas where residential areas are sharply differentiated in character.

Thus if an activity mainly employs a class of people dependent on public transportation to get to work (for example, very low-income people or married people whose spouses preempt the family car for their own commuting), it may have difficulty in recruiting an adequate work force in the less accessible suburban areas. Even if there is not an absolute shortage of applicants, there may not be enough of a surplus of applicants to allow much freedom of selection.

In the late 1950s, a study was made of the recruiting experiences of business firms in the Boston metropolitan area which had relocated to sites along a circumferential suburban freeway.34 Most of the establishments in the sample canvassed were sizable manufacturing plants, with electronics equipment the most numerous category. Every firm in the sample had made an advance survey of the residential and commuting patterns of its employees in an effort to anticipate any recruitment problems that the new location might involve. Several of the firms took explicit account of employee residential locations in choosing the specific section of the highway on which to relocate. Every firm found it necessary to establish a plant cafeteria at the new location.

The survey’s findings on recruitment problems were summed up as follows:

A summary of the observations of personnel managers interviewed indicated a general, but not universal, conclusion that Route 128 locations in comparison with the downtown areas definitely eased recruitment of engineering, professional and administrative staff, but neither helped nor hindered recruitment of skilled labor. Recruiting difficulties arose especially when the firms sought young, female clerical workers, male unskilled workers, and seasonal workers, both male and female, particularly in the higher income suburbs of the western subarea of Route 128. The type of recruitment problem mentioned most frequently was that of the younger, unmarried female, clerical workers.

The most serious of the recruitment problems was that of unskilled production labor, both male and female, but especially male, for seasonal work. Although this problem did not arise frequently, the need for a seasonal expansion of employment could cause a major headache for the personnel department, and [seems?] to indicate very sharply the lack of a casual labor market in the suburbs. At a downtown location, a firm could readily draw unskilled male and female workers with a "Help Wanted" sign in the window for seasonal employment. At a Route 128 location, this type of labor was scarce, and intown labor found commuting to the plant time-consuming and costly for low-paying jobs on a seasonal basis.

. . .The existence of a circumferential highway or of a few long-distance commuters does not lead to the conclusion that there is also a circumferential labor market, in which a firm at any one location can draw equally well from any other part of the area. On the contrary it would appear as if suburban firms tend to draw from areas nearby, and lose those workers who live at abnormally long distances away.

. . .Relationship of a firm to the immediate local labor supply seems to be crucial. Where the local supply is already committed, or of the wrong composition to meet the demands of the firm, the company will be forced to rely on longer-distance commuters. If this means extension of commuting beyond the normal range, or direction, for that type of labor, the firm will be likely to have major labor supply difficulties)35

It has been more than a quarter of a century since this report was published, but these findings are still relevant. The area around Route 128 (now an interstate highway) has become one of the nation’s major centers for advanced-technology activities; and as the metropolitan area has grown, the labor market in the vicinity of this highway has diversified substantially. However, the recruitment problems described in the report’s summary are now characteristic of firms considering locations on a yet more remote beltway I-495 around the Boston metropolitan area.

More recent and more elaborate studies have brought out some additional details concerning the characteristics of urban labor markets, For example, Albert Bees and George Shultz found in the Chicago labor market significant wage differentials for the same occupation among different neighborhoods within the metropolitan area, corresponding generally to the directions of commuter flow.36 Similar differentials have also been identified on the basis of distance from the central business district.37 The labor market of a large metropolis is clearly not a single spatially perfect market in the sense that location within it makes no difference.

As one might expect, the least mobile types of manpower (low-skilled workers, members of minority groups whose housing location choice is restricted by discrimination, and secondary and complementary workers)38 evince the greatest market imperfection in terms of differentials in their net earnings after deducting commuting costs. This same tendency toward wider geographical wage differentials for lower-income occupational groups has already been noted at the interregional level.

Workers are attracted in their residential choices toward job locations, whereas employers are attracted toward cheap labor supply. The relative force of these two sides of a mutual linkage varies widely among occupations, of course. At one extreme, the wage rate in an occupation is uniform throughout the labor market area, and the commuters from more distant residential areas bear all of the extra money and time costs of commuting over those greater distances. In the other extreme case, the employers pay higher wages at job locations farther from employee residential areas, absorbing the added costs of commuting longer distances by paying compensating differentials in wages.39

The amount of variation in wages in any given occupation will depend on the relative mobility of the employees and the employers and also on the extent to which union agreements or understandings among employers impose a single wage standard for the whole labor market area. It will depend also on two factors already mentioned; namely, the skill and income level of the occupation itself and the extent to which the workers’ residential choices are limited by discrimination.

It is clear that the spatial imperfection of large urban labor markets affects particularly the low-income worker. One of the most serious aspects of the present-day problem of inadequate job opportunities for residents of urban slums is that an increasing proportion of the kinds of jobs they might fill has shifted to distant suburban locations with little or no public transportation available, while very few employers have been willing to accept some obvious disadvantages of a slum location for the sake of closer access to that labor supply.


Several kinds of spatial differentials in earnings and income are of interest to various parties. The relative opulence of two communities can be compared in a marketing survey in terms of total or per capita money income. An individual looking for a good place to work would be interested in wage rates or annual earnings in his or her specific occupation, adjusted for any differences in the cost of living. An employer looking for low labor costs would want to compare specific wage scales, adjusted for productivity and fringe benefits.

Relative pay levels in specific occupations in the United States show a pattern of interregional differentials, with lower levels in the South and in smaller labor markets. These two differentials (North-South, and size of place) appear also in measures of per capita annual income and living costs. Income and pay levels deflated by cost-of-living indices seem also to show similar patterns, but to a much smaller degree. The term "equalizing differentials" is applied to a differential in money wages or income that merely compensates for a cost-of-living differential.

People move in response to perceived differences in prospective real incomes as well as other factors; migration flows depend on the characteristics of both the origin and the destination areas, the difficulties of the journey, and the characteristics of the migrant. Within a labor market area, housing and personal considerations account for most moves; for migration between labor market areas, job-related reasons are the most important.

The cost and difficulty of migration is roughly related to distance, and migration streams do show attenuation with distance, which can be expressed in a gravity-type migration model. Social distance is a broader term designed to take account of the degree of sociocultural adjustment required of the migrant in addition to costs of distance per se. Such considerations account for the beaten-path effect (migration is easier along paths used by many previous migrants from the same area).

Age is the personal characteristic most markedly related to migration rates, with generally declining rates to the late teens, a sharp peak around age eighteen to twenty, and a decelerating decline to old age with a small peak at retirement time. Migration rates are positively associated with both education and occupational status or skill, at any given ages.

Places with high rates of in-migration tend to have high rates of outward migration as well, because of the beaten-path effect and also because heavy in-migration produces a population with characteristics conducive to high mobility in terms of age, education, family status, disposition, and migration experience. Earnings levels and employment opportunity affect the amount of in-migration to a labor market but seem to have little effect on out-migration rates except in seriously distressed areas.

There has been a decline in overall migration rates in the United States since World War II. The factors contributing to this decline differ for short- and long-distance moves, but they include decreases in the average size of households, increased home ownership, and greater incidence of two-wage-earner households. Regions have also become more alike, thus reducing the incentive to migrate.

Labor mobility and the lack of it play a part in the development of labor cost differentials, which in turn affect the location of some interregionally footloose activities. Each such activity has its own preferences among labor supply locations, determined by the relative emphasis it places upon low wages, skill or trainability, and flexibility.

Low wages are most often found in relatively backward and/or depressed areas where labor demand has not kept up with the natural increase of the labor force. Limitations on outward mobility, particularly for the poor and less skilled, dam up in such places a pool of cheap surplus labor. In addition, low-cost complementary or secondary labor supplies occur in areas where there is a relatively heavy demand for the kind of labor services provided by the principal earners of families but relatively little demand for the services of other family members such as spouses or children.

For activities requiring a highly skilled or educated labor force, developed areas with high amenity are more likely to furnish the desired kind of labor supply. Flexibility and diversity of labor supply are most likely to be found in large metropolitan areas; however, these advantages are partially offset by more restrictive work and seniority rules.

Institutional constraints and the wage-bargaining procedures associated with large employee and employer organizations generally work in the direction of greater interregional wage uniformity in a given industry or occupation and also greater local uniformity among different industries and occupations within a single labor market area. There is some tendency for the wage spread between occupations to be wider in areas of slow employment growth or decline, reflecting in large part the higher mobility of the better-paid occupations.

Within large labor market areas, distance and commuting costs produce significant differentials in available labor supply, which are reflected in partially compensatory wage premiums at locations relatively far from workers’ residential areas.


Spatial mobility

Chain migration

Equalizing differentials in wages and incomes

Migration rate

Real differentials in wages and incomes

Selectivity of migration

Pull and push factors motivating migration

Positively and negatively selective migration

Social distance

Lowry hypothesis

Functional distance

Beale hypothesis

Beaten-path effect

Labor-oriented activities

Counterstream effect

Compensating wage differentials


Michael J. Geenwood, "Research on Internal Migration in the United States: A Survey," Journal of Economic Literature, 13, 2 (June 1975), 397-433.

Irving Hoch, "Income and City Size," Urban Studies, 9, 3 (October 1972), 229-328.

Chang-I Hua and Frank Porell, "A Critical Review of the Development of the Gravity Model," International Regional Science Review, 4, 2 (Winter 1979), 97-125.

Albert Rees and George P. Shultz, Workers and Wages in an Urban Labor Market (Chicago: University of Chicago Press, 1970).

Harry W. Richardson, Regional Economics (Urbana: University of Illinois Press, 1978), Chapter 5.

Larry A. Sjaastad, "The Costs and Returns of Human Migration," Journal of Political Economy, 70, 5 (2) (Supplement, October 1962), 80-93.



1. U.S. Department of Labor, Bureau of Labor Statistics, Handbook of Labor Statistics, Bulletin 2070 (Washington, D.C.: Government Printing Office, 1980), Table 120, pp. 292-293.

2. C. T. Haworth and C. W. Rasmussen, "Determinants of Metropolitan Cost of Living Variations," Southern Economic Journal, 40, 2 (October 1973), 183-192. See also Richard J. Cebula, "A Note on the Impact of Right-to-Work Laws on the Cost of Living in the United States," Urban Studies, 19, 2 (May 1982), 193-195.

3. The source for indices of consumer budget costs in selected metropolitan areas is given in Table 10-4. Data on per capita income for these and other metropolitan areas in 1980 can be found in U.S. Department of Commerce, Regional Economic Measurement Division, "Revised County and Metropolitan Area Personal Income," Survey of Current Business, 64, 4 (April 1982), Table 1, pp. 51-52.

4. The most widely cited study documenting the positive correlation of real income and population size in urban areas of the United States is Irving Hoch, "Income and City Size," Urban Studies, 9, 3 (October 1972), 299-328. A much earlier study of workers’ hourly earnings and living costs in Swedish industrial towns and cities showed a consistently positive relationship, with roughly twice as wide a range of variation in earnings as in living costs. The same study found consistently higher living costs in larger communities and in regions more distant from food-producing areas or having especially rigorous climates. See Bertil Ohlin, Interregional and International Trade (Cambridge, Mass.: Harvard University Press, 1933), pp. 215-218.

5. See Steven H. Sandell, "Women and the Economics of Family Migration," Review of Economics and Statistics, 59, 4 (November 1977), 406-414; and Jacob Mincer, "Family Migration Decisions," Journal of Political Economy, 86, 5 (October 1978), 749-773.

6. Of course, earnings and other benefits expected in the future must be properly discounted if valid comparisons are to be made. The seminal article recognizing this and other aspects of the individual’s decision to migrate as placing it among a number of human capital decisions, such as that to "invest" in education or on-the-job training, is Larry A. Sjaastad, "The Costs and Returns of Human Migration," Journal of Political Economy, 70, 5 (2) (Supplement, October 1962), 80-93. This view has since come to dominate empirical and theoretical attempts to explain migration on the basis of economic incentives. For an excellent survey of these and related developments—at least until 1975—see Michael J. Greenwood, "Research on Internal Migration in the United States: A Survey," Journal of Economic Literature, 13, 2 (June 1975), 397-433.

7. A number of recent studies have focused on quality-of-life factors or looked at the relationship between quality-of-life and economic factors in migration decisions. See, for example, Philip E. Graves, "Migration and Climate," Journal of Regional Science, 20, 2 (May 1980), 227-238; and Frank W. Porell, "Intermetropolitan Migration and the Quality of’ Life," Journal of Regional Science, 22, 2 (May 1982), 137-157.

8. Ohlin, Interregional Trade, p. 212. The term "equalizing" has sometimes been applied also to wage differences explainable by differences in the attractiveness of different occupations.

9. Separation of the push from the pull effect is obviously a tricky business, and there has been much controversy about how much weight to attach to each with respect to specific historical or current migrations. It has been argued that the migrations from European countries to the United States in the latter part of the nineteenth century were primarily motivated by pull, because they fluctuated from year to year in harmony with fluctuations in American business conditions but did not significantly vary in response to business cycles in specific European countries. See Richard A. Easterlin, "Long Swings in United States Demographic and Economic Growth: Some Findings on the Historical Pattern," Demography, 2 (1965), 497-500.

10. See Ira S. Lowry, Migration and Metropolitan Growth. Two Analytical Models (San Francisco: Chandler, 1966).

11. See Chang-I Hua and Frank Porell, "A Critical Review of the Development of the Gravity Model," International Regional Science Review, 4, 2 (Winter 1979), 97-125.

12. For a discussion of some issues relevant to the interpretation of the distance factor, see Aba Schwartz, "Interpreting the Effect of Distance on Migration," Journal of Political Economy, 81,5 (September/October 1973), 1153-1169.

13. H. ter Heide, "Migration Models and Their Significance for Population Forecasts", Milbank Memorial Fund Quarterly, 41, 1 (January 1963), 63-64.

14. Functional distance is a still broader concept, wrapping up in one index a measure of all factors impeding migration between a given pair of points. Functional distance can be evaluated by comparing actual migration flows with the flows that would he expected to occur simply on the basis of, say, the populations of the points in question. Functional distance can then be correlated with actual distance and other suspected determinants in order to break it down into components.

15. See for example, James B. Kau and C. F. Sirmans, "The Influence of Information Costs and Uncertainty on Migration: A Comparison of Migrant Types," Journal of Regional Science, 17, 1 (April 1977), 89-96, where the role of previous migrants as a source of information is examined statistically.

16. Lowell E. Galloway, Geographic Labor Mobility in the United States, 1.957 to 1960, U.S. Department of Health, Education, and Welfare, Social Security Administration, Office of Research and Statistics, Research Report No. 28 (Washington, D.C.: Government Printing Office, 1969). See also Janet R. Pack, "Determinants of Migration to Central Cities," Journal of Regional Science, 13, 2 (August 1973), 249-260.

17. E. G. Ravenstein, "The Laws of Migration," Journal of the Royal Statistical Society, 52 (June 1889), 241-301.

18. Arthur Redford, Labour Migration in England, 1800-1850 (Manchester: The University Press, 1926; 2nd ed., New York: A. M. Kelley, 1968). The same process was included as one of the basic principles of migration in Ravenstein’s analysis. It may be surmised that chain migration is less important than it once was in the United States, since the difficulties of moving almost certainly depend less than they once did on the sheer physical distance involved.

19. For example, see Table 28 beginning on page 60 of the source cited in Table 10-5 for data on metropolitan mobility by age and occupational group for the years 1975-1980). Note however that these and almost all other such tabulations record migrants’ characteristics as they were after the move. Many migrants move in conjunction with a change of occupation, and to that extent these figures are inaccurate in measuring migration probabilities of persons in particular occupation groups.

20. Peter M. Blau and Otis Dudley Duncan, The American Occupational Structure (New York: Wiley, 1967), pp. 271-272.

21. Harley L. Browning and Waltraut Feindt, "Selectivity of Migrants to a Metropolis in a Developing Country: A Mexican Case Study," Demography, 6, 4 (November 1969), 347-357.

22. Where the quality (of education, skills, income, or whatever) of a stream of migrants is, as in the Monterrey case, intermediate between that of the origin and the destination populations, the curious result is that the immediate effect of the migration is to lower the average quality in both areas, although at the same time it usually improves the quality of the migrants and of the populations of the areas combined! A similar apparent paradox is involved in the case of the legendary student who flunked out of Harvard and transferred to a rival New England institution (which shall remain nameless), raising the average scholastic level of both universities.

23 Everett S. Lee, "A Theory of Migration," Demography, 3, 1 (1966), 56. This article represents a thorough revision and amplification of Ravenstein’s much earlier "Laws of Migration.

24. See Lowry, Migration and Metropolitan Growth.

25. Calvin L. Beale, "Demographic and Social Considerations for U.S. Rural Economic Policy," American Journal of Agricultural Economics, 51, 2 (May 1969), 410-427. A fuller account of Beale’s findings appears in an unpublished paper, "The Relation of Gross Outmigration Rates to Net Migration" (presented at the meetings of the Population Association of America, Atlantic City, N. J., April 1969).

26. Lowry’s study was restricted to major metropolitan labor market areas.

27. U. S. Bureau of the Census, Current Population Reports, Series P-20, No. 377, Geographical Mobility: March 1980 to March 1981 (Washington, D.C.: Government Printing Office, 1983), Table A, p. 1.

28. On this and related questions, see Pittsburgh Regional Planning Association, Region in Transition, Economic Study of the Pittsburgh Region, vol. I (Pittsburgh: University of Pittsburgh Press, 1963), Chapter 4, especially pp. 106-109.

29. Ibid.

30. See Melvin W. Reder, "The Theory of Occupational Wage Differentials," American Economic Review, 45, 5 (December 1955), 846-847.

31. This relationship was noted earlier in connection with Table 10-1.

32. See E. M. Hoover, Location Theory and the Shoe and Leather Industries (Cambridge, Mass.: Harvard University Press, 1937), pp. 109, 215-217, and sources cited therein.

33. Robert Murray Haig, "Toward an Understanding of the Metropolis," Quarterly Journal of Economics, 40, 1 (February 1926), 195.

34. Everett J. Burtt, Jr., "Labor Supply Characteristics of Route 128 Firms," Research Report No. 1-1958 (Boston: Federal Reserve Bank of Boston, 1958; mimeographed).

35. The excerpts quoted here appear in a somewhat altered sequence.

36. Albert Rees and George P. Shultz, Workers and Wages in an Urban Labor Market (Chicago: University of Chicago Press, 1970).

37. See Randall W. Eberts and Timothy J. Gronberg, "Wage Gradients, Rent Gradients, and the Price Elasticity of Demand for Housing: An Empirical Investigation," Journal of Urban Economics, 12, 2 (September 1982), 168-176.

38. Secondary workers are those who are in the labor force intermittently or part time. Complementary workers (see Section 10.5.4) are those who belong to a household in which they are not the principal earner. There is of course a considerable overlap between these categories.

39. Rees and Shultz found substantial evidence of payment of part of the extra commuting costs of the more distant workers by large establishments in the Chicago labor market in the late 1960s. (This does not necessarily imply that specific employers paid wage premiums to those individual workers who had the longer commuting journeys, but rather that employers located long distances from residential areas of the type of workers they employed tended to pay higher wages than did employers located closer to such residential areas.) For a sample of workers in each of a number of occupations, Rees and Shultz found a positive correlation between wage rate and distance traveled to work when such other wage-influencing variables as the age, seniority, education, and race of the worker were also included in the regression equation. As a percentage of mean earnings, the extra compensation for added travel time ranged from 2 percent for janitors to 13 ½ percent for accountants. Rees and Shultz, Workers and Wages, pp. 169-175; see also Hoch, "Income and City Size," p. 316.

back to contents

back to previous chapter

next chapter