That leaves immigration, which has historically insulated the US from population decline and represents a kind of tap that the US can turn on and off. Over the next decade, it is set to become the primary driver of population growth for the first time in US history. The question now is exactly how much more immigration might be needed to accelerate population growth — and whether US policymakers can actually overcome their political differences on the issue to make it an effective tool. There are two main ways that the US could increase overall population growth: by encouraging people to have more children or by increasing immigration levels.
On their own, pro-natalist policies have historically failed to increase birthrates in the kinds of numbers that would be required to stave off stagnant population growth. Internationally, research has shown that child allowances have led to slight, short-lived bumps in birthrates.
From to , Spain had a child allowance that led to a temporary 3 percent increase in birthrates, but that was mostly because more people decided to have children earlier, rather than have more of them. After the allowance was revoked, the birthrate decreased 6 percent. Immigration is a much more reliable driver of population growth.
The average age of newly arriving immigrants is 31, which is more than seven years younger than the median American, meaning that they could help replace an aging workforce. They are also more entrepreneurial , which encourages economic dynamism, and more likely to work in essential industries , such as health care, transportation, construction, agriculture, and food processing. Immigrants may also help stave off regional population declines.
Immigrants are more likely to settle in areas where foreign-born populations already live, which are typically large metro areas that have lost population in recent years. Frey found in a report that, of the 91 large metro areas that gained population since the beginning of the decade, 15 would have actually lost population were it not for immigration, including New York, Chicago, Detroit, and Philadelphia.
In another 11 large metro areas, immigration accounted for more than half of their population growth. Refugees are also more likely to settle in less dense population centers where housing costs are lower, possibly reinvigorating the nearly 35 percent of rural counties in the US that have experienced significant population loss in recent decades.
To really reap the benefits of increased immigration, though, the US would have to ensure that immigrants have the ability to integrate, which it has done successfully in the past. Immigrants in the US already have a higher employment rate and labor participation rate than native-born citizens, and immigrant children tend to perform at or above the educational level of comparable US-born children.
In recent years, states and cities have adopted a patchwork of policies to promote immigrant integration, including programs designed to provide English classes, schooling, and professional training; resources to start businesses; and access to citizenship.
But Biden has reestablished an Obama-era Task Force on New Americans to expand the role of the federal government in such initiatives.
In recent history, before President Donald Trump pursued policies curbing immigration and global travel largely came to a halt during the pandemic, the US typically admitted more than 1 million immigrants per year. But under that scenario, census projections indicate that the US would see less than half the population growth between and than it saw over the previous 40 years. Some have argued that the US should try to set its immigration levels equal to its historical per capita rate of immigration, or to the per capita immigration rates of comparable countries, such as Australia or Canada.
We now examine available evidence for the first half of the s to see if the unique selectivity pattern of domestic net out-migration from high-immigration states persists for this period. It is not possible to undertake the detailed analysis of demographic subgroups for areas states or metropolitan areas that was conducted for the — Frey, b, c with decennial census migration data.
However, it is possible to compile reasonably comparable rates over the first four years of the s using the "migration one-year ago" question from the Census Bureau's Annual CPS. The rates for the — period can be compiled by adding the net migration components for each year and computing a rate based on the average mid-year population over the period. These rates, along with comparable rates for the — period from the census appear in Table The unique out-migration patterns shown for high-immigration states are generally apparent for both the late s and early s.
In most cases, there is a higher rate of net out-migration for persons with "less select" demographic attributes—those with less than a college education and those with incomes below poverty. Also, consistent with findings from the earlier period, selectivity is more pronounced for the white populations of these states than for the overall. Sample sizes preclude our conducting analyses specific to blacks or providing overall measures for Hispanics and Asians. The rates shown for the state of New Jersey provide an example.
Here, persons in poverty are most apt to leave the state. For example, in the early s, New Jersey's poverty population showed a net out-migration of Similar results were obtained when comparing the migration of persons with only high school educations or less than high school educations with those who are college graduates e.
It is useful to compare the selectivity patterns of California with those of Texas because, as mentioned above, these states underwent somewhat divergent economic circumstances between the late s and early s. That is, during the first period, California's economy was still relatively robust, while Texas was undergoing severe employment declines—conditions which reversed for the early s. Nonetheless, over both periods, each state's migrant selectivity patterns displayed an accentuated net out-migration for their poverty populations and either accentuated net out-migration or reduced net in-migration for persons with less education.
A more extensive analysis of this phenomenon for California can be found in Johnson and Lovelady, Indeed, during the "good" periods for each state — in California and — in Texas college graduates and nonpoverty persons were moving in while poverty persons were moving out.
This is consistent with the view that the poverty and unskilled segments of the population may be less responsive to the current cyclical conditions of the overall economy than they are to the labor competition and other out-migration-inducing pressures of immigrants to these states Frey, a. The general pattern of net out-migration shown in Table is unlike the "circulation of elites" characterization that is typically applied to interstate or intermetropolitan migration Frey, , b.
Usually, states that are losing migrants because they are undergoing economic downturns lose them disproportionately among their college graduate or more well-off segments of the younger population. In a like manner, states that are gaining internal migrants gain them disproportionately from these groups. The unique pattern of selective out-migration shown for most of these states during both the late s and early s is consistent with explanations discussed above that link immigration to domestic out-migration.
Although the previous descriptive statistics along with earlier analytic studies show a statistical relationship between immigration and the selective net out-migration of less-skilled native-born residents, no previous research has esti-. In this section, we present the results of such an impact analysis for states 48 contiguous states and the District of Columbia. Because the most recent detailed data, available for such an analysis, are based on the Census for migration over the — period based on the "residence 5-years ago" census question , we focus on this period.
Our analysis is restricted to examining the impact of different immigration levels on the domestic migration for persons with a high school education and less and who were aged 25—64 at the end of the migration period in The focus on this education attainment group is consistent with earlier research indicating that the unique selective out-migration response to immigration is largely confined to this group.
The focus on the age group 25—64 is for persons of labor force age who, for the most part, have completed their formal educations. In examining the impact of immigration on domestic migration, we focus on the changing levels of immigrants who are also less skilled high school education or less. This focus is consistent with the thesis that immigrants represent labor substitutes for domestic migrants with similar skill levels—an often-held explanation for the observed negative impact between immigration and domestic migration.
As such, our impact analyses will address two questions:. How would a 50 percent increase or a 50 percent decrease in current immigration of less-skilled labor-force-aged immigrants affect domestic migration patterns of less-skilled native-born Americans?
How would a similar increase or decrease in less-skilled immigrants to California only affect domestic migration patterns between California and other states? The answer to the first question would indicate the impacts of policies that would change the overall levels of immigration proportionately across skill levels, or those that would change the preference system in a way that would alter the numbers of less-skilled immigrants.
The answers to the second question are relevant to research findings that show that the domestic out-migration from California among less-skilled and poverty residents disproportionately relocates them to the nearby states of Washington, Oregon, Nevada, and Arizona Frey, a.
From the perspective of those states, domestic in-migration from California accounts for a large share of their overall in-migrating populations who have a high school education or less and incomes below poverty. This impact analysis will be able to assess the magnitude of California's domestic migration exchanges with these states which are affected by California's immigration levels. Methodological details for both components of this impact analysis are specified in Appendices A and B, respectively.
Our approach can be summarized in terms of two separate components. The first involves estimating the effects of low-skilled immigration on the migration process for native-born interstate migrants using the nested logit model that permits a separate estimation of residents' departures and migrants' destination choices as part of the overall migration stream process. This specific technique is one that has been developed by Liaw and his associates to examine migration processes in a number of contexts Liaw and Ledent, ; Liaw, ; Liaw and Otomo, ; Liaw and Frey, This analysis evaluates the effects on a state's domestic migration of low-skilled immigration to the state compared with other well-known migration determinants associated with the state's labor force, social and environmental amenities, and geographical contiguity.
The second component of the impact analysis uses the results of the nested logit model to generate estimated changes in migration rates, associated with assumed alternative low-skilled immigration levels, and applying these rates to appropriate populations at risk to generate various alternative interstate domestic migration outcomes that would be consistent with the assumed alternative immigration levels. Both components of this methodology are first estimated in separate age-disaggregated analyses for the broad age groups 25—29, 30—44, and 45—64 because these different age groups are subject to somewhat different mixes of migration determinants.
The results of these age-specific impact analyses are then aggregated to produce results for the entire 25to year-old age group for our population of interest i. Because the results of the nested logit model provide the basis for estimating the effect of low-skilled immigration on the domestic migration process, we summarize the results of these models here. Relevant equations for the depar-. The results for the destination choice submodel appear in appendix tables A4 , A5 , and A6 , for these respective age groups.
We note that our estimate of low-skilled immigration included in these models is based on — immigrants reported in the census, and, therefore, is likely to understate, to some degree, the number of illegal immigrants in this group. State variations in the cost of living, incorporating state variations in housing costs, are used to adjust the per capita income measure.
In addition to these, the destination choice model includes measures of distance and contiguity between origin and potential destination states. Definitions of all variables are listed in Appendix A. Within each age-specific 25—29, 30—44, 45—64 analysis, it is possible to interact the above state-level attributes with personal attributes including detailed age five-year age groups , race white, black, Asian, American Indian, Hispanic , education attainment below high school, high school graduate , poverty status below poverty, above poverty , and gender male, female.
This is because our analysis makes use of a detailed migration matrix that disaggregates — interstate moves by a cross classification of the demographic variables just described. This matrix was drawn from a special tabulation of the full ''long form" In estimating the final departure and destination choice models presented in Appendix A tables, a series of preliminary analyses were conducted to identify statistically significant interactions between state.
The measure of immigration used in this analysis identified all state residents with at most a high school education who reported a residence abroad in Although it would be preferable to employ a net immigration estimate comparable to the net internal migration estimate , neither the U. Census nor any other U. This use of the census "residence abroad" question is also consistent with previous research.
However, we note that this estimate does not necessarily overstate total net immigration to the United States, despite its omission of the emigration component. This is because migration from abroad, as reported in the census, substantially understates the illegal immigrant population. It is estimated that, during a given year, there is an emigration of between , and , residents of all education levels. However, it is also estimated that there is a net annual immigration of , illegal immigrants, many of whom are not counted by the census Martin and Midgley, Thus, the figures we use tend to overstate legal immigration but understate illegal immigration for the — period.
Because the latter is likely to be disproportionately comprised of those with lower education levels, our estimates of low-skilled immigration are understated.
Most of the effects in the final age-disaggregated nested logit models were consistent with expectations and can be found in Appendix A tables. Our main interest is in the impact of the low-skilled immigration rate on the departure rates of residents, destination choices of migrants, and interactions with personal characteristics, when other relevant attributes are controlled.
The low-skilled foreign immigration rate is defined on the basis of working-aged to year-old immigrants with high school educations or less 5 divided by the corresponding beginning-of-period state population. Our analysis indicates that the primary impact of low-skilled immigration on native-born migration operates through the departure from high-immigration states, rather than as a reduced tendency to choose such states as destinations.
This is because the contribution to total explanation, associated with low-skilled immigration, is much stronger in the departure models than it is in the destination choice models.
In fact, low-skilled immigration has an almost similar effect in the departure models as do the combined effects of the labor market variables.
Its contribution to explanation is stronger than the combined labor market variables for persons in the 45—64 age group, suggesting that many of these pre-retirees are influenced as much by factors associated with states with high immigration levels as by standard income, unemployment rate, and employment growth attributes.
Equally noteworthy as the strong impact that low-skilled immigration exerts on the departure of low-skilled domestic residents is its interaction with specific subgroups. Strong interactions are shown for whites and especially whites below the poverty line. This is consistent with descriptive analyses that indicate, when controlled for education, that poverty residents are most likely to leave high-immigration states Frey, c; Frey et al.
Other significant interactions with low-skilled immigration are shown for blacks, poor blacks, and poor Hispanics see appendix tables A1 , A2 , and A3.
Although low-skilled immigration is not an important explanatory factor in the destination choices of migrants, it is noteworthy that the racial similarity of a destination state shows as much explanatory power as the conventional labor market variables see appendix tables A4 , A5 , and A6. This is especially the case for blacks, Hispanics, and Asians as well as Hispanics with less than a high school education and consistent with earlier observations that longer-term residents and native-born members of the new immigrant minority groups are likely to locate in areas with large numbers of same-nationality residents.
In assessing alternative immigration levels for impact analysis, we have chosen to focus on the age groups 15—64 because alternative immigration policies are likely to affect the entire labor- force-aged population.
Hence, although the focus of our domestic migration impact estimations are persons aged 25—64 for reasons discussed in the text , we assume that their migration patterns will be affected by changes in low-skilled immigrant levels at all labor force ages.
In sum, the nested logit model analyses confirm the results of earlier research, indicating that states with high levels of immigration of persons who might be labor substitutes for resident workers will show an accentuated out-migration of lower-skilled native residents when other state attributes are controlled.
This effect operates more strongly through the departure part of the migration process than through the destination choice. Our results also confirm earlier findings suggesting that the effect is most prominent among lower-income native-born residents of these high-immigration states.
The results of these models are incorporated into the impact analysis discussed in Appendix B. Our first set of impact analyses makes two alternative assumptions.
The first assumption is that the observed level of immigration for working-aged immigrants with at most a high school education is decreased by , over the — period. The second assumption is that the current level is increased by a similar amount over the — period. These numbers approximate 50 percent increases, or 50 percent decreases, of such immigration compared with the observed levels in the census.
These increases and decreases occur proportionately to each state with respect to their actual immigration levels. For convenience, we refer to these assumptions as 50 percent increases in immigration and 50 percent decreases in immigration. The analyses below present the estimated impacts that these assumptions imply for net domestic migration of states' native-born residents, aged 25—64, with high school educations or less.
The results of these scenarios for each state are shown in Table These data make plain that when immigration is decreased, it is the high-immigration states that tend to retain more of their native-born domestic low-skilled migrants who might otherwise have relocated to a low-immigration state.
Similarly, when immigration is increased, high-immigration states are the most prone to lose domestic native-born low-skilled migrants to other states. This is depicted in Map which shows the estimated change in net domestic migration for states on the assumption that there is a 50 percent increase in U.
Under this scenario, only ten states including the District of Columbia would show increased net domestic out-migration or in the case of Florida reduced net domestic in-migration with at most a high school education.
These changes would accrue to California , , New York , , Florida , , Texas , , and Illinois , The states gaining most from these net domestic migration shifts would be Arizona 36, , Georgia 22, , Pennsylvania 22, , and Nevada 22, Clearly, the states surrounding California, and those in the South Atlantic region—presumably attracting migrants from New York, New Jersey, Massachusetts, Illinois, and Florida— would gain larger numbers of less-skilled domestic migrants under this scenario of higher immigration.
MAP Estimated five-year change in net domestic migration for states assuming 50 percent increase in U. Figure shows how the two different assumptions about immigration would impact the net domestic migration for four high-immigration states: California, New York, Texas, and Illinois.
California would show the greatest disparities in the net domestic migration of their less-educated population. Given the observed level of immigration over the — period, California shows an expected domestic out-migration of 59, low-skilled residents. However, if immigration were reduced by 50 percent, the state would show a gain of 44, low-skilled persons, while under a 50 percent increase immigration scenario, it would lose , such persons. The other three states show net domestic out-migration under each scenario with New York showing the widest fluctuation of these three.
Another way to assess the impact of these different immigration scenarios is to examine how many low-skilled domestic migrants a state would gain for every low-skilled immigrants who did not come in under a reduced immigration scenario ; or to estimate how many low-skilled domestic migrants it would lose. These figures are shown for states with greatest immigration in Table , where the changes are calculated on the basis of immigrants and domestic migrants aged 15— Results show that there are not similar levels of exchange under both the decreased immigration and the increased immigration scenarios.
For example, under a decreased immigration scenario, California would gain 27 low-skilled domestic migrants for every such immigrants who did not come into the state. Yet under an increased immigration scenario, California would lose 51 low-skilled domestic migrants for every additional such immigrants that came into the state. Migrant exchanges under the increased immigration scenario are the most dramatic: The exchange of low-skilled immigrants for low-skilled domestic migrants is 2 to 1 in California.
It is just as strong in New York and Illinois, and about 5 to 2 in Texas. Because our impact analysis permitted a disaggregation by race-ethnicity, poverty status, and detailed age, it is possible to examine the impact of immigration changes for different demographic groups of a state's population.
Figure presents an analysis for California, based on our simulation, that shows the domestic net migration rates specific to different groups under the assumptions a that actual immigration levels occurred over the — period. As in the above analysis, these results pertain to persons with a high school or less education and also are restricted to the ages 25— The results show that under the conditions of actual immigration levels, rates of net out-migration are somewhat higher for low-skilled whites than for the low-skilled population overall.
The out-migration rates are substantially larger among the low-skilled poverty population, and especially the white low-skilled poverty population. When a 50 percent reduction in national immigration levels is assumed, these patterns change noticeably. Under the latter scenario, the net migration. However, the most dramatic changes occur with the two poverty groups.
The rate for the total low-skilled poverty population changes from Even more significantly, the rate for the white low-skilled poverty population changes from to These results indicate that the impact of immigration on the net domestic out-migration of California's low-skilled population is somewhat more accentuated among whites but that it is especially important in accounting for the net out-migration of the low-skilled poverty population.
These results also hold up when the analysis is confined to specific age groups so that they do not reflect patterns of only the younger or older populations within the state. Moreover, similar simulations with other high-immigration states show that, as with California, immigration disproportionately affects their low-skilled white and poverty populations see Liaw et al.
These impact analyses have shown that reduced or increased levels of low-skilled immigration show considerable effects on the redistribution of less-skilled domestic migrants for high-immigration states. The fact that, under an increased immigration scenario, new low-skilled immigrants to California will precipitate a net out-migration of 51 low-skilled native migrants from California suggests that there is a substantial demographic displacement occurring in this state.
Another important finding of this analysis is the "spillover effects" that changing immigration levels impose indirectly on low-immigration states as a result of increased or decreased domestic migration out of high-immigration states. Under the scenario of a 50 percent increase in immigration nationwide, 39 states would register increased net domestic migration gains on low-skilled native-born residents.
Alternatively, most of these states would lose low-skilled domestic migrants to the high-immigration states under a scenario of a 50 percent reduction in immigration to the United States.
Following this discussion of "spillover effects" of immigration, we now focus on an impact analysis that assumes that only California experiences a 50 percent decrease or 50 percent increase in its low-skilled immigration levels. The purpose of this, as indicated above, is to assess the indirect impacts of these changes on California's domestic migration exchanges with nearby states: Washington, Oregon, Nevada, and Arizona.
From a numeric standpoint, we are assuming under a "decreased immigration scenario" that California's — immigration of low-skilled labor-force-aged persons is reduced by , Similarly, under the "increased immigration" scenario, we assume that an additional , such immigrants move into the state. The "spillover" impact of these assumed reductions and increases in California's immigration can be seen in Figure Shown here are the net domestic gains of low-skilled native-born migrants for the States of Washington, Oregon, Nevada, and Arizona under three different California immigration scenarios.
It is clear from these statistics that Arizona and Nevada show the greatest changes as a result of these different scenarios. If immigration to California were reduced by 50 percent. Arizona's net domestic gains of 69, low-skilled migrants would become reduced to 44, If California's immigration level were to increase by 50 percent, Nevada's net domestic gains of low-skilled migrants would be raised from 38, to 68, Smaller, but similar, fluctuations are observed for Oregon and Washington.
The ''spillover effects" of immigration to California on surrounding states are selective on different demographic groups within the low-skilled populations just discussed. To illustrate this, we present results from our simulations for Nevada and Arizona that compare their net domestic migration patterns under the conditions in which a California received its actual immigration levels over the — period, and b California's immigration levels were reduced by 50 percent over the — period.
The comparison for Nevada is shown in Figure and indicates that when California's immigration levels are not re-. However, when California's immigration is reduced, it has the spillover effect of reducing, disproportionately, domestic net migration of the low-skilled poverty population to Nevada.
The latter scenario has the effect of reducing Nevada's domestic migration gains for all of the groups shown in Figure , but the impact is especially large on Nevada's poverty population. A similar impact is also shown for Arizona in Figure Here, a reduction in California's immigration levels has a disproportionate effect on Arizona's net domestic migration levels for both its poverty population and its white poverty population. Under the assumption of actual immigration to California, Arizona' low-skilled domestic migration gains are somewhat higher for whites than overall, but are substantially higher for low-skilled whites in poverty.
Under the assumption of reduced immigration to California, Arizona's low-skilled poverty migration rates are dramatically lowered, such that for poor whites, the domestic. These results for Nevada and Arizona indicate that the spillover effects of immigration to California disproportionately impact on the low-skilled poverty populations in these neighboring states.
This overview of the impact of recent immigration on population redistribution within the United States has shown that there is a continued concentration of immigrants to selected port-of-entry states and metropolitan areas at the same time the redistribution of internal migrants is more dispersed. Our own studies and those of others suggest that the concentration of immigrants is, in part, a function of their proclivity to locate in areas where there are existing concentrations of persons with like race-ethnic backgrounds and nationalities, and that.
In contrast, the internal migration processes, at work over the late s and early s, are more responsive to conventional labor market "pushes" and "pulls" and are drawn to state and metropolitan area destinations that are not the major immigrant port-of-entry areas.
However, there is clear evidence of accentuated domestic out-migration from high-immigration states and metropolitan areas for persons with high school educations or less. The multivariate analyses conducted for this study indicate that this is associated with low-skilled immigration, after controlling for relevant labor market and amenity variables that are generally used to explain interlabor market migration. Moreover, our impact analyses suggest that the immigration impacts on this domestic out-migration are considerable and would approach a 2-to-1 relationship in California under a scenario of a 50 percent increase in the state's recent immigration level.
The explanation for this demographic displacement may well lie with arguments that immigrants represent labor substitutes for domestic migrants who can take advantage of opportunities in other areas, However, our results are also consistent with other explanations as well. For example, less well-off, longer-term residents in high-immigration areas may be reacting to perceived increases in social costs that may take the form of higher crime rates, reduced services, or increased local taxes that they may take to be a function of recent immigrant flows.
In addition, one cannot ignore the possibility that race and ethnic prejudice may enter into decisions of native residents, especially whites, to relocate away from increasingly multiethnic areas in much the same manner that such prejudice prompted "suburban flight" in many American cities in the s and s.
Moreover, our findings suggest broader implications for changes in the social demography of high-immigration areas if recent immigration and internal migration patterns persist. For example, it has been argued that port-of-entry metro areas are taking on a "dual economy" character where large numbers of immigrants, participating in lower-skilled and informal sectors of the labor force, provide complementary activities for more advanced services and corporate headquarters activities among the mostly white-native professional ranks Mollenkopf and Castells, ; Sassen, ; Waldinger, The demographic implications of this scenario become apparent when examining the foreign-born shares and minority shares of different socioeconomic attributes in high-immigration metros Table For these metropolitan areas, the foreign-born population comprises a disproportionate share of persons without high school diplomas, in the lower quartile of family income, and of workers in service and unskilled blue collar occupations.
The imbalance is even more pronounced in the Los Angeles metropolitan area where, for example, foreign-born residents comprise three-fifths of all persons whose family incomes fall in the bottom quartile, while representing only 23 percent of those in the upper quartile. Over half of service.
The divergence in the nativity-class structure for the combined high-immigration metro areas and individual areas, such as Los Angeles and New York, contrasts markedly with the rest of the United States—where the foreign born comprise only 6 percent of persons aged 18 and above and disparities by socio-economic measures are not nearly as skewed. More contrasts can be made with respect to the minority composition of high-immigration metros and the rest of the United States Table , right panel and on other demographic attributes typically associated with the foreign-born population.
The statistics for point up already sharp disparities with respect to the class-nativity and class-race-ethnic structures between the metropolitan regions that serve as ports of entry and other parts of the United States. The findings in this chapter suggest that there is a continued concentration of immigration associated with race-ethnicity, country-of-origin groups, and those with high school educations or less that is occurring at the same time that internal migration is redistributing longer-term, native-born migrants to different metropolitan areas and to smaller communities as well as nonmetropolitan territory.
The latter movement is, to a large degree, a function of more traditional labor market pushes and pulls, as well as amenities, that are not centered in the same port-of-entry areas in which immigrants are concentrating.
However, a significant part of the native-born movement away from high-immigration metropolitan areas, among residents with a high school education or less, has been shown to be related to the levels of recent immigration in those areas when other factors are controlled.
Thus, immigration would appear to exert both direct and indirect impacts on redistribution within the United States as a result of its concentration in selected areas and its secondary impact on net domestic out-migration from those areas.
To the extent that it is the lower-skilled immigrants with at most a high school education who are most associated with both patterns, our results suggest that immigration policies that would select more highly educated immigrants might serve to alleviate immigrant concentration as well as the selective demographic displacement of the native born by immigrants that has been observed over the — period.
Beyond the immediate implications that these concentration and selective displacement patterns hold for local economies and the employment options for less-skilled native-born workers, our findings also suggest that there are broader implications associated with the changing social demographics of high-immigration areas and with the likely widening demographic disparities between areas of high immigration and other parts of the country.
More bifurcated race-class labor force structures, the changing demographic profiles of child poverty populations, and widening race-ethnic disparities across state populations are just a few of the social demographic consequences that can result from a con-.
The emerging social demographic patterns hold important implications for the nation's social and political geography and deserve further examination and study as consequences of current immigration policies. A two-level nested logit model of interstate migration is applied in this study to the — interstate migration data to assess the effects of low-skilled immigration on U. The estimated results of this model are an integral part of the impact analysis that is discussed in further detail in Appendix B and in the relevant section of the text.
A useful feature of the nested logit model as applied to the current problem is its ability to identify separately the determinants of residents' departure from a state and those of migrants' destination choices.
The first section below discusses the general methodology of this model. The second section presents the variables that are used to construct the model. The final section consists of Tables A1 through A6 that present the estimated coefficients for the departure submodel and destination choice submodel, respectively, for the specific age groups 25—29, 30—44, 45—64 as background for the discussion of this model's results in the text. These models are termed the "best models" because, on the basis of extensive preliminary analyses, they represent those in which variables retain sensible and statistically significant coefficients.
The formulation of the two-level logit model is as follows. The demographic attributes include detailed age 25—29, 30—34, …, 60—64 , race white, black, Asian, Hispanic, American Indian , educational attainment below high school, high school graduation , poverty status below or above the poverty line , and gender. Our choice of potentially useful explanatory variables to be considered for inclusion in the model is guided by 1 previous research findings in the literature, 2 the hypotheses we wish to test, and 3 extensive preliminary cross tabulations of the migration data.
In constructing a relatively comprehensive model to be called the best model for simplicity , we include only the explanatory variables that are statistically significant i. The data sources for these variables are described in Frey et al. Note that the ceiling of Rho-square is much less than 1. The Hispanic population, 8 million in , is nearly 57 million in The Asian population, 1. The white population of nearly million grew to million in The black population, 21 million in , increased to 40 million in Over the next five decades, the majority of U.
By , the United States will be a nation without a majority racial or ethnic group. The Hispanic and Asian populations will continue to grow more rapidly than whites and blacks. Among immigrants, the white share of the foreign-born population will remain at a historic low over the next five decades.
Based on Pew Research Center assumptions about immigration, fertility and mortality rates, Asians are projected to surpass Hispanics as the largest single group among the foreign-born population, beginning in The white and black shares are expected to rise over the next five decades, while the Hispanic and Asian shares will decline somewhat.
Today, Asians are the only major racial or ethnic group whose numbers are rising mainly because of immigration. Although immigration contributes to growth of the U. Hispanic population, births in the U. Births overtook immigration as the main driver of Latino population growth after Krogstad and Lopez, In , immigration rates had been low for several decades, so both immigrants and their U. The contrast is striking in terms of median age—the age at which half the population is older and half younger.
For the U. The median age for immigrants was The median age for the second generation, whose parents mainly came during the early 20th-century wave of immigration, was By , the U. Had there been no immigration after , the population would have been slightly older, with a median age of nearly The most striking change, though, is in the second generation.
0コメント