The mass migration of African Americans out of the South during the first two-thirds of the twentieth century represents one of the most significant internal migration flows in U.S. history. Those undertaking the Great Migration left the South in search of a better life, and their move transformed the cultural, social, and political dynamics of African American life specifically and U.S. society more generally. Recent research offers conflicting evidence regarding the migrants’ success in translating their geographic mobility into economic mobility. Due in part to the lack of a large body of longitudinal data, almost all studies of the Great Migration have focused on the migrants themselves, usually over short periods of their working lives. Using longitudinally linked census data, we take a broader view, investigating the long-term economic and social effects of the Great Migration on the migrants’ children. Our results reveal modest but statistically significant advantages in education, income, and poverty status for the African American children of the Great Migration relative to the children of southerners who remained in the South. In contrast, second-generation white migrants experienced few benefits from migrating relative to southern or northern stayers.
Introduction: The Great Migration
At the turn of the twentieth century, the southern African American population had ample reason to consider leaving the South. Decades of discrimination and oppression following the Civil War had resulted in educational deprivation, economic disadvantage, political weakness, and thousands of victims of lethal southern mobs (Anderson 1989; Beck 2015; Blackmon 2008; Kousser 1974; Mandle 1978, 1992; Margo 1990; Tolnay and Beck 1992). In 1900, the vast majority of southern-born blacks continued to reside in southern states, with less than 5 % located outside the South (Tolnay 1999, 2003). In contrast, southern-born whites were migrating north and west in substantially larger numbers than their black counterparts.
By the mid-twentieth century, however, the situation had changed dramatically. Roughly 20 % of southern-born blacks lived outside their region of birth, with an overwhelming share of those migrants located in a handful of northern and western cities such as Chicago, Los Angeles, New York, Detroit, and Philadelphia (White et al. 2005). By contrast, roughly 15 % of southern-born whites resided outside the South and were distributed more widely across states, cities, and even nonmetropolitan areas (Tolnay 1999; White et al. 2005). This massive stream of black and white migrants from the South persisted from roughly 1915 through 1970 and is commonly referred to as the “Great Migration.” The substantial regional and residential population redistribution that resulted from the Great Migration altered the economies, political landscapes, racial climates, and cultures of both the South and the non-South (Gregory 2005).1
How did the southern migrants and their families fare in “the promised land”? Classical migration theories assume that, on average, the type of interregional movement that characterized the Great Migration should benefit the migrants (Borjas 1987; Greenwood 2015; Lee 1966; Ravenstein 1885; Ritchey 1976). Many have investigated the socioeconomic outcomes for southern out-migrants, assessing whether this interregional movement did, in fact, benefit those who moved. Our study is the first to use longitudinal data to reveal how the Great Migration influenced the socioeconomic outcomes of the migrants’ children—the Great Migration’s second generation—in their adulthood. Our study includes both white and black migrants. We compare these migrant groups with second-generation stayers, whose parents never left the South, as well as with those who were born to northern-born parents and residing in the non-South.
The consideration of second-generation and even third-generation migrants is quite common in the literature on international migration (Alba 2005; Boyd 2009; Boyd and Grieco 1998; Chiswick and DebBurman 2004; Farley and Alba 2002; Kalmijn 1996; Park and Myers 2010; Portes et al. 2009; Reitz et al. 2011; Sakamoto et al. 2010; Thomas 2012; Trejo 2003). This expanded focus is motivated by the recognition that first-generation migrants who cross national borders face significant adjustment challenges—for example, language and cultural differences—that will prove less daunting for their children and grandchildren. It is also based on the recognition that first-generation international migrants typically seek better lives for themselves and their broader family networks, including their children. In many ways, the gulf between southern and nonsouthern states during much of the Great Migration was similar in scale to the cross-national contrasts encountered by international migrants. As a result, generational variation in the adjustment process—and ultimate well-being—among migrants is likely. Furthermore, those who left the South surely considered the benefits of their decision for their children, especially for African American migrants who sought both socioeconomic opportunity and social justice.
Indeed, racial differences are an important part of the story of the Great Migration and, by extension, our examination of the experiences of second-generation migrants. Whites and blacks alike experienced the consequences of limited economic opportunity in the South, including retarded industrial development and an agricultural system based heavily on the labor of landless farm tenants and sharecroppers (Mandle 1978, 1992; Ransom and Sutch 1977; Ruef 2014). Southern blacks were confronted with the additional burden of Jim Crow laws and customs as well as frequent racially motivated violence (McAdam 1982; Morris 1984; Tolnay and Beck 1995; Woodward 1951). In a similar vein, the contexts of reception in the North and West for black and white migrants from the South were not identical, with occupational opportunities highly stratified by race (Lieberson 1978) and with many northerners viewing increases in the black population with particular hostility (Drake and Cayton 1945; Ignatiev 1995; Roediger 1991; Spear 1967; Trotter 1985). In light of these racialized differences in the sending and receiving regions of the Great Migration, it is reasonable to anticipate racial variation in the experiences of the migrants.
In this study, we analyze the long-term outcomes for the Great Migration’s second generation using novel longitudinal data constructed by linking individuals in the 1940 census to the 2000 census long form. These data allow us to observe the location of individuals in 1940 and 2000 as well as the location of their own and parents’ birthplaces. From this information, we identify the children of Great Migration participants and determine whether these children were born before or after their parents left the South. We also observe the socioeconomic characteristics of each individual’s childhood home in addition to their own adult outcomes. Our results reveal modest but statistically significant advantages in education, income, and poverty status for African American children of the Great Migration relative to the children of southerners who did not leave the South. However, second-generation white migrants largely did not differ from their southern or northern stayer counterparts after the characteristics of their parents in 1940 are accounted for.
Outcomes in the Great Migration
The question of whether migrants benefitted from leaving the South is relatively recent. For years, researchers and commentators assumed that individuals did benefit socially, politically, and economically, given the significant “push” factors from their origin location, the “pull” factors to their destination locations, and the logic of rational choice theories of human migration (e.g., Greenwood 2015; Lee 1966; Massey et al. 1993; Ritchey 1976). Letters written by migrants and potential migrants, as well as oral histories, revealed powerful economic, social, and political incentives for blacks to leave the South (Faulkner et al. 1982; Scott 1919; Wilkerson 2010). Likewise, black as well as white southern-born migrants were motivated to move north and west by myriad factors, including the mechanization of agriculture, the decimation of cotton crops by the boll weevil infestation, widespread farm tenantry, and the intensification of cash crops (Berry 2000; Gregory 2005; Mandle 1978). Given the assumptions intrinsic to traditional migration theory, as well as the testimony of the migrants themselves, it seems self-evident that those who left the South reaped significant economic and social benefits. The reality, however, is more complicated.
In many respects, southern black migrants to the north and west tended to have better outcomes than their northern-born counterparts. Black southern migrants exhibited lower rates of separation and divorce (Tolnay 1997, 1998; Tolnay and Crowder 1999; Wilson 2001), were more likely to be employed (Tolnay 2001), spent more of their working lives in full-time jobs (Lieberson and Wilkinson 1976), and earned higher incomes (particularly among young males and those who resided in the North or West for a few years) (Lieberson 1978; Long and Heltman 1975; Masters 1972). In contrast, white migrants tended to fare slightly worse than native Northerners in terms of income, poverty status, and occupational attainment.
These advantages for black southern migrants relative to the native nonsouthern population were due, in part, to the selectivity of southern migration to the North and the West. Research has suggested that the migrants were positively selected in terms of education (Black et al. 2015; Gregory 2005; Margo 1988; Tolnay 1998); income (Black et al. 2015); occupational prestige (Collins and Wanamaker 2014); and, in some periods, marital status (Mahoney 2007; Wilson 2001). Likewise, although the majority of southern blacks during this period were employed in the agricultural sector, the majority of black southern migrants held nonfarm occupations (Collins 2000; Marks 1989), and many migrated from towns or cities with established railway lines, indicating that these migrants were primarily urban (Alexander 1998; Black et al. 2015; Wilkerson 2010). More recently, Boustan (2016) examined the occupations of migrants’ fathers in 1920 and concluded that the selection of migrants from the South actually was bimodal: stronger among those from lower and higher economic origins than among those in the middle.
Despite the selectivity of northward black migrants and their apparent success relative to northern-born blacks, recent evidence is mixed regarding how both black and white migrants fared relative to their stationary counterparts in the South. Compared with southern-born black individuals who did not move, or who moved within the South, interregional black southern migrants exhibited lower life expectancy (Black et al. 2015), were more likely to be incarcerated (Muller 2012), were no more likely to participate in entrepreneurial activities or to be employed in most professional positions (Boyd 2012, 2013), were more likely to be unemployed, and had lower (or not substantively different) relative incomes (Eichenlaub et al. 2010; Epstein 1919). The outcomes for white southern-born migrants have been studied less extensively. However, Eichenlaub et al. (2010) found, with a few exceptions, that white southern-born migrants to the North did not fare better than their southern stayer counterparts or migrants within the South in terms of employment status and income, and they fared worse in terms of relative income and occupational status.
Boustan (2016) reached a different conclusion regarding the comparative economic standing of black southern migrants in the North and West and their counterparts who remained in the South. Comparing brothers residing in the same household in 1935, only one of whom left the region by 1940, Boustan concluded that the migrants enjoyed substantial economic benefits from leaving the South—roughly doubling their annual income, adjusting for regional differences in the cost of housing (i.e., rents). Collins and Wanamaker (2014) also concluded that migrants earned more than their sedentary southern counterparts, using a linked 1910–1930 census sample of black males and an indirect estimate of earnings based on industry of employment, occupation, and region of residence.
A great deal of complexity remains behind the general questions of whether southern migrants fared better economically than did native northerners or their stationary counterparts in the South. The Great Migration took place over a half-century, a span of time during which both the sending and receiving regions changed appreciably. Southern agricultural production modernized, while its industrial and manufacturing base expanded. The Civil Rights Movement and black activism changed the racial climate throughout the nation but especially in the South (Andrews 2004; McAdam 1982; Morris 1984). The labor needs of northern industry evolved, even as employment opportunities and populations gradually migrated from central cities to suburbs. The magnitude of the Great Migration, itself, also may have shaped post-migration outcomes for those leaving the South as destination labor markets became more competitive and origin labor markets became less competitive. Indeed, Boustan (2009, 2016) found that the increased numbers of black workers in the non-South late in the Great Migration depressed blacks’ wages while leaving wages for northern whites unchanged. At the same time, it is possible that stationary southern blacks potentially saw their incomes rise as the population of blacks in their region declined. Transformations in the southern contexts from which the migrants departed, as well as in the northern contexts in which they arrived, may therefore have influenced the structure of opportunities available in the North and the South and thereby influenced the fate of the migrant population.
Significantly, the vast majority of research on the relative success of southern migrants during the Great Migration has been based on the experiences of the first-generation movers—those who exited the South in favor of the North and West. Relatively little attention has been devoted to assessing the experiences of the children of those migrants—that is, the second generation.
Second-Generation Migrant Outcomes
The international migration literature may help to situate the surprising result that first-generation migrants may not have enjoyed the level of economic rewards predicted by classic migration theory. In general, this literature reveals that first-generation immigrants tend to have poorer economic outcomes, whereas the second-generation benefits from the moves of their parents. Specifically, second-generation immigrants obtain higher levels of education (Boyd and Grieco 1998; Farley and Alba 2002; Park and Myers 2010; Reitz et al. 2011; Sakamoto et al. 2010; Thomas 2012), income (Kalmijn 1996; Reitz et al. 2011; Sakamoto et al. 2010; Trejo 2003), homeownership (Park and Myers 2010), and occupational status (Boyd 2009; Boyd and Grieco 1998; Chiswick and DebBurman 2004; Kalmijn 1996; Reitz et al. 2011), and experience lower rates of poverty (Thomas 2011) relative to first-generation immigrants and, in most cases, their native, third- or higher-generation counterparts. Moreover, these results are generally consistent for Hispanic and Latino, Asian, Caribbean, African, and European immigrant groups, although the degree of occupational and economic adaptation2 and upward mobility varies in magnitude by race and ethnicity.
Muted economic success for first-generation migrants during the Great Migration could cohere with the larger narrative concerning international immigration and generational advancement. International immigrants and black southern-born migrants during the Great Migration share many similarities, including the tendency for both groups of migrants to be positively selected (Dodoo 1997; Dustman et al. 2012; Feliciano 2005; Kusow et al. 2016; Min and Jang 2015). Moreover, despite this positive selection, the returns to education and training acquired in the place of origin of both groups of first-generation migrants are less than expected (Akresh 2006; Boustan 2016; Dodoo 1997).
While migrants during the Great Migration obtained their education and job experience in the United States, northern labor markets may have placed less value on human capital gained in the South. Indeed, southern-born migrants were popularly characterized as backward and uneducated (Berry 2000; Gregory 1989; Lemann 1991; Wilkerson 2010) despite the fact that many migrants resided in relatively urban locations and held higher than average socioeconomic statuses prior to leaving the South. Additionally, the southern educational system was widely known for its lack of quality and short school year, particularly for African Americans (Anderson 1989; Faulkner et al. 1982; Margo 1990; Marks 1989; U.S. Bureau of Education 1916).
In contrast, most children of first-generation southern migrants obtained their education and work experience in their nonsouthern destination locations. Similarly, the positive selection of their parents likely translated into a more stimulating than average home environment, and parents may transmit to their children the aspirations for mobility and achievement that motivated them to move in the first place, thereby inculcating a drive to succeed. Any or all of these qualities may influence the outcomes of second-generation migrants, potentially allowing them to reap the rewards brought about by their parents’ migration and to advance beyond their stationary counterparts in the South. Because of the dearth of longitudinal data collected during the period of the Great Migration, the experience of second-generation migrants has gone relatively unexplored. Those studies that have examined this question provide evidence generally supportive of second-generation success. Tolnay and Bailey (2006) found that second-generation female migrants from the South claimed the highest probability of school enrollment during the early twentieth century, exceeding that of black, white, and Jewish female adolescents from all other generational groups. Likewise, Tolnay (2001) found that second-generation migrants from the South obtained higher occupational statuses relative to both first-generation migrants from the South and native northerners in 1920. Restifo et al. (2013) also observed that second-generation black southern-born migrants obtain higher status occupations relative to first- and third-generation African Americans, and these generational advances, although marginal, were actually greater than the generational advances made by many white immigrant groups.3 Therefore, suggestive evidence exists that second-generation black southern-born migrants benefitted socioeconomically from the moves of their parents.
Data and Methodology
We use newly constructed data linking individuals from the 1940 census to the 2000 census long form.4 The second-generation migrants at the core of our study are children in the 1940 census who were living with a parent who had moved from the South. We establish parent-child relationships via the 1940 census “relationship to household head” response: the head, spouse, and children were linked as a family. The use of the 1940 census allows us to observe characteristics of children’s parents and household, including whether one or both parents were present. Because we focus on second-generation migrants who could be linked to the 2000 census, the average individual in our analysis is between 65 and 70 years of age in 2000.5 Our outcome variables, therefore, describe the second-generation migrants in their later adulthood.
The linkage of the 1940 and 2000 censuses relied on the record linkage infrastructure developed within the U.S. Census Bureau Center for Administrative Records Research and Applications (CARRA). The 1940 and 2000 censuses were both processed using a set of programs and tools—the Person Identification Validation System (PVS)—which appends unique Protected Identification Keys (PIKs) to census records and other files. PIKs can be used to link records between any files that have been processed using PVS (Wagner and Layne 2014).6 Because the PVS employs administrative data that records name changes before and after marriage, we are able to accurately link women from their childhood observation in 1940 to their adult observation in 2000.
There were 38,235,897 children aged 0–18 enumerated in the 1940 census, of whom 27,107,415 (70.9 %) were assigned a PIK necessary for linkage to other PIKed files (Alexander et al. 2014). Of these, 3,169,843 could be linked via PIKs to the 2000 census long form.7 The primary factors working against a successful linkage between an individual’s 1940 and 2000 census records are (1) death or migration to another country between 1940 and 2000; (2) non-enumeration, or enumeration with error, in the 2000 census; (3) incorrect PIK assignment in the 1940 or 2000 censuses; or (4) failed PIK identification in the 2000 census (85.5 % of cases in the 2000 census were assigned PIKs). We compare the characteristics of all children in the 1940 census with those who received a PIK and were linkable to the 2000 census and find that they are similar (see Table S1, Online Resource 1). We also conduct all the analyses reported later herein while reweighting the linked 1940–2000 sample to be representative of black and white children in the 1940 census; the results are quantitatively similar, suggesting that the linkage process introduces little bias.8
We limit our sample to children 18 years or younger who lived with one or both parents in the 1940 census and who could be linked to the 2000 census. Among those cases, we further limit the sample to children who were enumerated as black or white, were born in the United States, have reported income and education in 2000, and whose reported race is consistent between 1940 and 2000.9 These sample restrictions reduced the number of usable linked cases to 2,268,156.
Children in our sample were grouped into one of four migration status categories: southern stayers, southern-born migrants, northern-born migrants, and northern stayers.10 This classification was made according to the following criteria:
|Southern stayers11||Parent birthplace: South (either parent)|
|Child birthplace: South|
|Residence in 1940: South|
|Migrants, southern-born||Parent birthplace: South (either parent)|
|Child birthplace: South|
|Residence in 1940: non-South|
|Migrants, northern-born||Parent birthplace: South (either parent)|
|Child birthplace: non-South|
|Residence in 1940: non-South|
|Northern stayers||Parent birthplace: non-South (both parents)|
|Child birthplace: non-South|
|Residence in 1940: non-South|
To assess the relationships between second-generation migration status and socioeconomic outcomes, we conduct bivariate and multivariate analyses separately by race, which allows us to relax the assumption that the experiences of black and white second-generation migrants were identical, consistent with the substantial racial differences in the societal contexts in the sending and receiving regions mentioned earlier.
Table 1 reports the number of observations within each migrant and racial group. Of the 123,528 black individuals in the sample, there are 102,190 southern stayers and 4,126 northern stayers. Members of these two groups did not have a parent who migrated from the South to the North or West.
Of the black, second-generation Great Migration migrants, we observe 2,592 southern-born second-generation migrants and 14,620 northern-born second-generation migrants. Among whites, we observe 603,174 southern stayers, 1,415,904 northern stayers, 21,453 southern-born second-generation migrants, and 104,097 northern-born second-generation migrants.
The set of outcomes that we compare across migration status groups includes (1) the probability of graduating from high school, (2) logged total personal income for the year 1999,12 and (3) the probability of living below the poverty line. In supplementary analyses, we also compare the number of years of education obtained by each migrant status group.
We use ordinary least squares (OLS) regressions to examine differences in the logged value of total personal income in 1999 across migrant groups and, in supplementary analyses, the number of years of education obtained across migrant groups. We use logistic regression analysis for the binary outcome variables graduating from high school and living below the poverty line.
Our sample results pool men and women. We originally examined the results separately for males and females. However, we found that the relationships between migration status and socioeconomic outcomes were very similar in direction and magnitude.
We include sets of control variables drawn from both the 1940 and 2000 censuses. For 1940, when the second-generation resided in their parental homes, we include residence in rural or metro areas, farm status, residence in one of the top five southern migrant–receiving cities,13 the presence of two parents (vs. one parent), the highest grade achieved by either parent, and the highest occupational status exhibited by either parent.14 Based on the respondent’s characteristics in 2000, when we observe the respondent in late adulthood, we control for age and squared age, residence in noncentral city or central city metropolitan statistical areas (MSA), and gender. For the analyses of income and poverty status, we also control for years of education and state median income in addition to MSA status to account for regional differences in the cost of living.
The control variables all serve to identify as closely as possible the “uncontaminated” differences in education, income, and poverty status across the migration status groups, but the substantive heterogeneity among the control variables should be noted. Some are intended to account for simple compositional differences across migration status groups that might be related to the outcome variables: for example, gender, age, and MSA and central city residence in 2000. Others, such as parental education and occupational status, have the potential to adjust for selectivity of the southern migrants. Finally, some have the potential to tap the importance of early life conditions or experiences for influencing later-life outcomes, such as residence in a two-parent family or on a farm or in a metropolitan area in 1940. However, given data limitations, these functional differences in covariates cannot be as distinct as one would like. For instance, parental education and occupational status also could have been shaped by post-migration opportunity structures available to the first-generation migrants as well as by pre-migration conditions in the South. Even residence in one of the top five migrant-receiving cities could represent an indirect pathway through which exiting the South affected socioeconomic standing in older adulthood. These inherent ambiguities should be kept in mind as the following empirical evidence is presented and discussed.
We present the descriptive statistics for all outcome and predictor variables used in the analyses, by race and migration status, in Table 2. Even in the raw descriptive statistics, the data reveal differences in the socioeconomic outcomes of second-generation migrants relative to stayers. Parental education and occupational attainment scores are higher for migrants relative to southern stayers. Moreover, in adulthood, black migrants exhibit more years of schooling and higher probabilities of graduating from high school, higher incomes, and lower probabilities of being in poverty.
The findings from our empirical analyses are presented graphically in Figs. 1, 2 and 3. The light gray bars represent the predicted scores obtained from the bivariate analyses, and the darker gray bars represent the predicted scores obtained from multivariate models with all covariates (except for median state income, which is excluded from these models) held constant at their mean values. In the models for income and poverty status, the darkest gray (almost black) bars represent the findings from multivariate models that control for all the aforementioned covariates, including median state income. The full tables of results for each of these models can be viewed in Online Resource 1 (Tables S2–S5).
Panel a in Fig. 1 reveals that second-generation black migrants and northern stayers exhibit substantially higher probabilities of graduating from high school than southern stayers. Specifically, in bivariate models, southern-born and northern-born second-generation migrants are roughly 30 % and 48 % more likely to graduate from high school, respectively, than are southern stayers. Second-generation migrants continue to experience superior educational outcomes relative to southern stayers after the control variables representing both 1940 and 2000 characteristics are added to the model. However, the magnitude of these advantages decreases substantially, as can be seen in the comparisons between the predicted values for models with and without covariates in panel a. After we control for individual, familial, and parental characteristics, southern- and northern-born black second-generation migrants are 11 % to 12 % more likely to graduate from high school than southern stayers—still a substantial albeit considerably reduced gain over southern stayers. Additionally, second-generation migrants do not exhibit significantly different predicted probabilities of graduating from high school relative to northern stayers.
The reduced differences in high school completion by migration status are due largely to the inclusion of the 1940 control variables. The 1940 covariates, including residence in rural or metro areas, parental educational and occupational attainment, and homeownership, all have important and distinct influences on these relationships and help to account for the relationships between migration status and the probability of graduating from high school. Some of these 1940 covariates likely reflect the positive selection of migrants from the South. Therefore, the importance of parental characteristics in accounting for differences in high school graduation probabilities by migration status among the second generation could point to the selectivity of migrants and the importance of their relatively high levels of urbanization and educational and occupational attainment and relatively low levels of rural residence for their children’s success in adulthood. As noted earlier, however, this same pattern of findings, suggestive of a positive selection effect, also could be the result of the first-generation southern migrants taking advantage of greater post-migration social, economic, and residential opportunities in the North and West.
These same characteristics may also help to explain why southern-born second-generation migrants tended to have lower probabilities of obtaining a high school diploma in the bivariate models relative to northern-born second-generation migrants and northern stayers. After the 1940 covariates are included, these differences disappear, perhaps because southern-born second-generation migrants spent more time in rural areas and may have had less time to adapt. For example, southern-born second-generation migrants likely had less opportunity to take advantage of the superior northern educational system than their northern-born counterparts. Thus, the lower educational attainment of southern-born second-generation migrants can likely be traced to their early life experiences, the experiences of their parents, and the consequences of these experiences for their occupational and economic adaptation into their northern environments.
For whites (Fig. 1, panel b), the educational benefits of migrating are more ambiguous. In bivariate analyses, southern-born second-generation migrants are only 4 % more likely to graduate from high school than southern stayers, but northern-born second generation migrants and northern stayers are approximately 15 % more likely to graduate from high school. However, after we control for relevant covariates—particularly, the parental 1940 characteristics—these differences in educational attainment among whites largely disappear, and all groups exhibit probabilities of graduating from high school that range from 0.81 to 0.83. Southern-born second-generation migrants no longer exhibit significantly different probabilities of graduating from high school relative to southern stayers; northern-born second-generation migrants are only 2 % more likely to graduate from high school, and northern stayers are 2.7 % more likely to graduate from high school. Once again, parental characteristics explain much of the narrowing of these differences for both black and white second-generation migrants and northern stayers.15
As with the findings for high school graduation, black second-generation migrants report significantly higher levels of income than their southern stayer counterparts (Fig. 2, panel a). In the bivariate analyses, southern-born second-generation migrants and northern-born second-generation migrants earn predicted annual incomes that are approximately $2,784 and $4,839 higher than that of southern stayers, respectively—a gain in income of 22 % and 38 %. After including the full set of covariates, except for median state income, these differences narrow to $980 and $1,164, respectively. Finally, including median state income narrows these differences further to $617 for southern-born second-generation migrants and $793 for northern-born second-generation migrants relative to southern stayers, representing a 5 % to 6 % gain in income over southern stayers. Moreover, both migrant groups achieve levels of income that are not significantly different from northern stayers, again indicating their ability to successfully adapt to their northern location.
The income benefits of leaving the South are considerably smaller for whites (Fig. 2, panel b), echoing the findings for high school graduation. In bivariate analyses, both second-generation migrant groups and northern stayers earn higher incomes than southern stayers. After accounting for all 1940 and 2000 control variables, however, these income differentials fall and become slightly negative. Indeed, after accounting for median state income and the full set of covariates, white second-generation migrants earn 4.4 % to 6.3 % less than southern stayers. However, both migrant groups earn slightly more than northern stayers. As with education, it is the characteristics of black and white second-generation migrants’ parents that largely explain any advantage they have over southern stayers.
When interpreting the differences in reported income for the year 1999 across migration status groups, it is important to keep in mind the very substantial changes that had occurred in the southern and nonsouthern economies during the previous 60 years. During this time period, many areas within the South experienced significant industrial expansion and agricultural modernization. At the same time, major northern urban areas suffered from deindustrialization and central city decline (Massey and Denton 1993; Wilson 1978). As a result of both these trends, the southern exodus was reversed as both blacks and whites abandoned the Rust Belt in favor of the Sun Belt (Falk et al. 2004; Frey 2004). It is very possible that the incomes of second-generation Great Migrants would have compared more favorably with those for southern stayers in, say, 1959 than they did in 1999.
The black second-generation migrant advantage persists when the focus turns to the probability of being in poverty, as might be anticipated from the evidence for income (Fig. 3, panel a). Among black southern stayers, this probability is quite high. Indeed, in bivariate models, southern stayers have a 0.189 probability of being in poverty, compared with a 0.127, 0.116, and 0.117 probability for southern-born second-generation migrants, northern-born second-generation migrants, and northern stayers respectively.
After all covariates, except for median state income, are included in the models, these differentials narrow but do not disappear. Southern-born black second-generation migrants are still approximately 15 % less likely than their southern counterparts to be in poverty (0.136 compared with 0.161), and northern-born second-generation migrants and northern stayers are approximately 12 % less likely to be in poverty than southern stayers. Including median state income narrows these differentials further so that southern-born and northern-born second-generation migrants are approximately 11 % and 7.1 % less likely to be in poverty, respectively.
As with education and income, the relationship between migration status and the probability of being in poverty is partially explained by the 1940 covariates. Thus, part of the advantage of belonging to one of these migrant groups can be explained by the higher levels of education and occupational status obtained by second-generation migrants’ parents. Moreover, second-generation migrants and northern stayers are more likely than southern stayers to have parents who resided in metro areas in 1940 and who, predictably, were less likely to reside in rural areas in 1940. Again, although these early familial characteristics hint at the selectivity of the migrants, they may also reflect the relatively greater post-migration opportunities available to those who left the South versus those who remained behind.
White migrants and northern stayers also experience lower probabilities of being in poverty than southern stayers, although these differences are not large given the low probability of being in poverty among white respondents in general (Fig. 3, panel b). The differences, although small in magnitude, indicate that second-generation migrants’ probability of being in poverty is approximately one-fourth to one-fifth lower than the probability for southern stayers. After we include the full set of covariates but exclude median state income, these advantages shrink dramatically, although the differences remain statistically significant. However, after we include median state income, neither second-generation migrant group exhibits predicted poverty levels that are significantly different from that observed for southern stayers. In contrast, northern stayers are approximately 8 % less likely to be in poverty than sedentary southerners.
Discussion and Conclusion
Research on the socioeconomic outcomes for participants in the Great Migration has focused nearly exclusively on the outcomes experienced by the migrants themselves—that is, first-generation migrants. However, as with most international migration, the decision to leave the South was often meant not only to improve the migrants’ own life chances but also to improve the life chances of their children—second-generation migrants. We found mixed evidence that these goals were achieved.
Our results demonstrate that migration to the North and West yielded advantages for migrants’ children, advantages that are apparent far into their adulthood and are particularly evident among black migrants. Specifically, black second-generation migrants exhibit higher incomes and probabilities of graduating from high school and lower probabilities of being in poverty than their southern stayer counterparts. Moreover, for all outcomes, the second-generation migrants did not fare worse, and sometimes even fared better, than northern stayers. Thus, their occupational and economic adaptation into the North appears successful.
However, these advantages can be explained largely by the characteristics of migrants’ and northern stayers’ parents. Both migrant groups and northern stayers tend to have parents with higher levels of education and occupational status and lower probabilities of residing in rural areas in 1940 than their southern stayer counterparts (generally consistent with Boustan’s conclusions (2016)). As noted earlier, the importance of the 1940 characteristics for accounting for variation in outcomes across migration status groups could reflect the operation of diverse processes. First, many of these characteristics could themselves have been influenced by the migration experience. All the migrant groups moved by 1940, and the characteristics of parents and families were measured in 1940. Thus, parents could have higher occupational statuses precisely because they moved. Families could also be much more likely to live in a metropolitan area because they moved. Another possible explanation for the importance of the 1940 characteristics is that they indicate selectivity in the migration of parents from the South. Specifically, perhaps the more highly skilled, educated parents with a taste for urban life are those most likely to move. These explanations are not mutually exclusive and could be operating simultaneously.
In an attempt to adjudicate between these potential explanations, we sequentially added parental educational attainment in 1940 into the multivariate models, followed by the remaining 1940 characteristics. Because parents likely completed their education prior to moving, this factor is less likely (although not irrefutably) to be influenced by the post-migration experience. Thus, if parental education explains away the relationship between migration status and the socioeconomic outcomes, it would point to the importance of selectivity. However, if the other 1940 characteristics are more important for explaining the relationships, it would point to the importance of post-migration experiences for shaping early life experiences. We found that parental educational attainment in 1940 explained much of the relationship between migration status and our socioeconomic outcomes. The other 1940 characteristics also explained sizable portions of the relationship, but their mediating effect was not as large as the mediating effect of parental education (results available upon request). This would suggest that the selectivity of parental migration is an important factor influencing the later-life outcomes of second-generation migrants, although it is likely that the post-migration experience also shaped other elements of respondents’ early-life experiences—such as parents’ occupational statuses—that in turn positively influenced later-life outcomes. Thus, migrating to the North and West yielded some benefits for black migrants’ children, and these advantages endured even after we accounted for a wide variety of individual and familial characteristics. Nevertheless, these benefits were smaller than one might anticipate, given contemporary descriptions of “the promised land,” and may perhaps indicate that the non-South was a less hospitable environment for many migrants than is often assumed.
For white respondents, few benefits were associated with exiting the South after individual and familial characteristics, particularly the 1940 characteristics, were controlled. Thus, the advantages associated with belonging to either of these second-generation migrant groups largely appear to be due to the selectivity of parental migration and the consequent importance of this selectivity for the second generation’s childhood experiences and later-life outcomes.
In focusing on whether southern migrants fared better socioeconomically, we do not wish to obscure the diverse social, cultural, and political benefits that were associated with moving to the North or West, such as the presumably lower levels of overt racism and racially motivated violence, higher levels of political enfranchisement, and generally weaker legal supports for racial segregation and discrimination (Gregory 2005; Tolnay and Beck 1992; Wilkerson 2010). We therefore acknowledge the importance of recognizing that many southern black and white migrants experienced improvements in their political and social environments, improvements that may be difficult or impossible to measure with the data at hand, even as they might have experienced modest socioeconomic benefits. Nevertheless, it is important to determine whether migrants or their children benefitted socioeconomically from their moves because socioeconomic advancement itself provides access to a wide array of other societal advantages.
These findings are unique and represent an important contribution to the literature on the Great Migration and the returns to migration more broadly. This is the first study that we know of to link Great Migration migrants to their children, and no other study we know of has been able to examine the benefits of internal U.S. migration for second-generation migrants into their late adulthood. We are therefore able to bring the insights of the international immigration literature, which has often emphasized the benefits for migrants’ children, to the case of internal migration within the United States.
Second, our finding that the Great Migration marginally benefitted black migrants’ children is important in light of previous research findings that the migration experience may not have benefitted the migrants themselves. We also demonstrated that the benefit of migration for second-generation migrants can be explained largely by the characteristics of their parents, offering a unique opportunity to examine the influence of early-life experiences on later-life outcomes. Moreover, our study did so without relying on retrospective reporting and the attendant problems in recall.
Additionally, our inclusion of female migrants in our study is an important contribution because female migrants have often been neglected in the migration literature, and relatively few studies of the Great Migration have accounted for the experiences of female migrants (see, however, White 2005). Furthermore, few studies linking individuals over time have been able to do so for females, largely because women often take their husbands’ names after marriage, therefore making such linkages very challenging.
Finally, our study directly compared the benefits and costs of migration for black and white migrants. In doing so, we were able to show that the advantages associated with migration were considerably larger and more consistent for blacks and that whites seem to have experienced advantaged socioeconomic positions relative to blacks that are persistent across migrant and nonmigrant experiences. That second-generation black migrants compared favorably with their sedentary counterparts in the South while white migrants did not could speak to the greater barriers to upward social mobility encountered by southern blacks throughout much of the twentieth century, at least until the de jure foundation for Jim Crow discrimination and segregation was successfully dismantled by the Civil Rights Movement.
This study is subject to limitations. We were unable to assign PIKs to approximately 30 % of children in 1940 and were unable to locate 30 % of those in the 2000 census who were assigned a PIK in 1940. It is possible that the ability to link individuals across censuses is correlated with our outcomes. The fact that we were unable to link 100 % of the children in the census also affects the external validity of our sample; however, when we reweighted the linked sample to be representative of all black and white children aged 0–18 in 1940, the results remained largely the same.
Moreover, our sample was limited to older individuals: those who were, on average, aged 60–70 in 2000. Our results are therefore not generalizable to younger, more recent groups of second-generation migrants—for example, those whose families participated in the heavy out-migration from the South during and after World War II.
Although our data offered unprecedented opportunities to explore the well-being of second-generation migrants from the Great Migration, data limitations restricted our ability to estimate the relative contributions of migrant selectivity and post-migration opportunity to the observed variation in socioeconomic status by migration history. Therefore, our findings should be considered largely descriptive, and causal inferences about the benefit of leaving the South accruing to second-generation migrants must be drawn with caution.
Our study provides valuable answers about the fate of second-generation southern migrants, but it also raises a number of questions that can help to chart the direction for future research on this topic. First, we recommend that the racial differences yielded by our analyses be further interrogated. Perhaps more specific answers can be identified for why the benefits of leaving the South were greater for blacks than for whites. Second, more work is required to adjudicate between migration selectivity and post-migration opportunity as forces responsible for any advantaged outcomes inferred for migrants over nonmigrants. Third, more attention needs to be devoted to how the dramatic socioeconomic, political, and cultural changes occurring in the South and non-South throughout the twentieth century influenced the relative well-being of those who left the South versus those who remained, and how those contrasts changed over time. Finally, we encourage scholars to expand future inquiry to include possible beneficial migration outcomes beyond those examined in our study—for example, greater political power, improved civil liberties, and enhanced cultural expression opportunities (Gregory 2005). Clearly, there is great potential for future scholarship on the Great Migration and the fate of those who participated.
Partial support for this research came from awards from the Eunice Kennedy Shriver National Institute of Child Health and Human Development to the Center for Studies in Demography and Ecology (CSDE) at the University of Washington in support of its training program (T32 HD007543) and infrastructure (R24 HD042828), as well as from the CSDE Shanahan Endowment Fellowship. We would like to thank the participants in the 2016 American Sociological Association, particularly Jenna Nobles, for their valuable comments and suggestions. We are also grateful for the feedback from the five anonymous reviewers during the Demography peer review process. This research was conducted as a part of the Census Longitudinal Infrastructure Project (CLIP) while J. Trent Alexander was an employee of the U.S. Census Bureau. Any opinions and conclusions expressed herein are those of the authors and do not necessarily represent the views of the U.S. Census Bureau.
We occasionally refer to the “South” and “non-South,” using the latter to refer to all states outside of the census-defined South. In some cases, we distinguish between the “North” and “West” as distinct nonsouthern locations. When we use the generic terms of “North” and “northern,” without additional distinction, we are referring to all areas outside the South.
We borrow this term from Portes and Rumbaut (1990).
The study by Restifo et al. (2013) is based on public use microdata samples for 1910, 1920, and 1930 for New York City only. They defined the second-generation as those born to one southern-born parent, and the third-generation as those born to two northern-born parents. Their data imposed a number of important limitations, including (1) a lack of early life characteristics for second-generation migrants and (2) an inability to clearly delineate when individuals migrated. Further, females and southern-born whites are not included in their study.
The Minnesota Population Center and Ancestry.com (2013) provided the complete-count 1940 census.
We did not extend our analysis to the 2001–2015 American Community Surveys (ACS) because we believed this would exaggerate selection into the matched sample that occurs by requiring survival between 1940 and 2000. Longevity is positively associated with higher education outcomes (Lleras-Muney 2005) and income (Chetty et al. 2016). Given the age of our sample, the relationship among education, income, and mortality may lead to our estimating upper bounds for the benefits experienced by second-generation migrants.
Because the 2000 census long form includes roughly one in six U.S. households, the maximum number of children with PIKs from 1940 who could be linked to their 2000 census records is 4,526,938 (i.e., 0.167 × 27,107,415 = 4,526,938). In principle, we successfully linked roughly 70 % (i.e., 3,169,843 / 4,526,938 = .70) of the children from the 1940 census who had PIKs assigned. Because we do not account for migration or death in these estimates, a match rate of 70 % is a worst-case scenario, or lower bound, on how many individuals we should link from 1940 to 2000.
We follow a reweighting procedure similar to that proposed by DiNardo et al. (1996) to reweight the matched sample to exhibit the distribution of characteristics of the full population. We construct the weights using probit regressions of matched status on characteristics in 1940, including migrant group, age, parental education, parental occupational standing, and whether the respondent lives with one or both parents (results available by request). We also include polynomials in age, parent education, and parent occupation score up to the fourth degree. Due to computational constraints, we took a 10 % random sample of the 31 million total black and white children in the United States to reweight against. Our probit regressions estimate the propensity of being matched or unmatched (p). We then reweight the matched cases by (p / (1 – p) × (1 – q) / q), where q is the share of matched cases. Only three point estimates are statistically different from their original values as a result of the reweighting exercise. These include the coefficients on white northern stayers for the probability of graduating high school (without covariates), the coefficients on whites who migrated before birth and white northern stayers in the income regressions (without covariates), and the coefficients on migrated before birth in the probability of being in poverty regressions (without covariates) for whites.
Approximately 0.5 % of the total cases have nonmatching race responses in 1940 and 2000. Because this is not a sufficient number of cases to support a “both” racial category in the analysis, we exclude the cases.
We also examined whether migrants who moved to the West differed from those who moved to the North (Midwest and Northeast). However, we found few significant differences between these groups in our multivariate models, and so those comparisons are not examined here.
We estimated our models while both including and excluding southerners who migrated out of the South between 1940 and 2000 from the group of southern stayers. However, those results yielded very similar results. In fact, our decision to include in our primary analyses individuals who may have migrated out of the South between 1940 and 2000 slightly attenuates the advantages enjoyed by black migrants.
Income includes wages, salaries, commissions, bonuses, tips, self-employment income, public assistance, Social Security income, and Supplemental Security Income.
We control for the top five destination cities for both black and white migrant groups to control for network effects caused by previous migration streams as well as location-specific differences in income and employment opportunities. The top five cities for blacks are Detroit, Chicago, St. Louis, Philadelphia, and New York City. The top five cities for whites are Chicago, Los Angeles, Cincinnati, Detroit, and Cleveland.
We measure occupational status using occupation scores constructed using median income by occupation from the 1950 census (Sobek 1995).
We also examine the relationship between migration status and educational outcomes using OLS regression, with years of educational attainment as the dependent variable. We find that the returns to migration are somewhat narrower for black second-generation migrants using this specification. However, the results are substantively very similar to the models examining blacks’ and whites’ probabilities of graduating from high school, with second-generation black migrants exhibiting modestly higher years of educational attainment relative to southern stayers, and white second generation migrants exhibiting statistically nonsignificant differences in educational attainment relative to white southern stayers after we control for the 1940 characteristics. We present these results in Table S5 and Fig. S1 in Online Resource 1.