Companion Piece to Number 4 in Series, "Opportunity in America"
The views expressed are those of the author(s) and do not necessarily reflect those of the Urban Institute, its board, its sponsors, or other authors in the series.
The authors would like to thank Robert Hauser, Michael Hout, and Seymour Martin Lipset for very helpful comments on an earlier draft of this paper without in any way implicating them in our final conclusions.
Young men today have lower incomes than their fathers' generation did at a similar age. 1 Further, many measures of income inequality in the United States have been growing for over twenty years. Both of these phenomena raise significant questions about a broader issue: the extent of social and economic opportunities available in today's society. Generational income comparisons or statistics on the extent of income inequality, however, do not provide sufficient evidence to make an accurate judgment on the crucial question of how much opportunity exists for an individual growing up in the United States. These aggregate data must be complemented by an understanding of the extent to which a child in today's society is able to do better than his or her parents, and whether this is related to a more open structure in which class matters less than it used to or simply to an expansion of the total opportunities available to each generation as the result of economic growth. This paper examines the relationship between an individual's socioeconomic status and the status of his or her parents (or intergenerational mobility) and the reasons for any change in this relationship over time.
What Determines Intergenerational Mobility?
Intergenerational mobility for any one individual is determined primarily by two factors: (1) the amount of opportunity in society, and (2) the rate of economic growth and associated change in the occupational structure. 2 These two components of mobility are discussed in additional detail below. 3
The Role of Opportunity
Opportunity, in this context, is defined as the degree to which income and social status are determined by the innate skills and ambitions of an individual, and not by any inherited advantages or disadvantages. The more closely an individual's socioeconomic status is determined by parents' status, the less opportunity exists. The more independent the two are, the more opportunity is present, and the more likely it is that children with equal abilities will have an equal chance to succeed. 4
Intuitively, the concept of opportunity is simple enough. A society with a significant degree of opportunity would be one in which children reaped little or no advantage from being born into a rich family rather than a poor one. 5 Those with the most skills and ambition, regardless of family background, would be the most likely to achieve success. Other factors such as social class, race, and gender would not affect occupation or earnings.
On the other hand, a society with little opportunity would be one in which individuals are not allowed to rise or fall on their own merits. A caste system, in which everyone is assigned a role at birth, is an example of such a society. Regardless of their individual talents, most children would be expected to end up in the same occupations as their parents and have similar living standards, reflecting the lack of individual opportunity.
The Role of Economic Growth
The rate of economic growth also fuels intergenerational mobility because productivity growth is the fundamental factor that drives wages (and, to a lesser extent, per capita living standards). Over time, improvements in overall productivity and corresponding increases in wage levels tend to make children, on average, better off than their parents were. In an economy with ever-increasing productivity, children may surpass their parents' standard of living simply through overall improvements in living standards even if there is little or no movement between the ranks of the rich and poor in that society. Thus, it would be possible to have a society in which there is no individual opportunity, but in which every child still climbs the economic ladder.
Throughout American history, the majority of children have been able to do better than their parents. This is primarily the result of economic growth although differential fertility rates by social class can also have an effect. 6 Still, in the absence of economic growth, movements up and down the economic ladder tend to offset each other.
How Much Opportunity Would We Expect in an Open Society?
In the United States, our public philosophy is grounded in the idea of opportunity. Americans believe more strongly than individuals in other countries that success should and does reflect individual effort. 7 But it should be recognized that a significant degree of intergenerational immobility will remain, even in a system that is open, fair, and dynamic. This is, in part, a result of biological and family-related influences that may have the effect of perpetuating class advantages or disadvantages.
Genetic inheritance alone would lead us to expect at least some relationship between the positions of parents and their children, no matter how much opportunity is available in a society. If income-earning ability is even partially influenced by genetic characteristics, there will be a positive relationship between parents' and children's earnings, regardless of how equally opportunity is provided in that society. One oft-cited factor that may affect income-earning ability is an individual's measured cognitive ability, or "intelligence." We estimate, however, that not more than 4 percent of the variation in children's earnings can be associated with the genetic transmission of cognitive ability. 8 The actual proportion is likely to be smaller than even this tiny fraction, for our estimate of the heritability of intelligence includes (to an unknown extent) the influence of environmental, as well as genetic, factors. 9
Beyond genetic transmission, of course, there will always be other factors at work that have the effect of strengthening the intergenerational link. There will likely always be a natural tendency for parents who occupy positions of high status to try to extend their privileges to their children. Indeed, Lipset and Bendix (1959, 2) note that "a 'good' father is one who tries to pass the status he enjoys on to his children...Hence, in every stratified, complex society there is, as Plato suggested, a straining toward aristocracy and a limitation of mobility." Public institutions can partially compensate for this tendency, but are unlikely to be able to compensate fully short of imposing extraordinary measures that would drastically disrupt the rights of families to raise their own children.
Thus, it is difficult to determine the maximum degree of opportunity that would be possible in a society in which some fraction of talents and ambitions are passed to the next generation through various biological and familial processes. In addition to the numerical estimate discussed above, we can also gain some insight into the level of mobility in American society, however, by comparing mobility in the United States to that in other countries. 10 Such an exercise enables us to determine more accurately whether this country provides more opportunity to its citizens than is available in other nations. Further, if the extent of intergenerational immobility falls outside the range discussed above, then we might conclude that the playing field in the United States is not entirely level, and that class or family background continues to make a difference. We will make use of both of these forms of comparison in the sections that follow.
A Note on Terminology
In the academic literature, the two factors discussed above (opportunity and economic growth) are referred to as circulation mobility and structural mobility, respectively. Circulation mobility (sometimes called exchange mobility) is the independent association between (a) an individual's occupation or socioeconomic status (often referred to as the individual's "destination"), and (b) the occupation or socioeconomic status of the individual's parents at an earlier time (the individual's "origin"). The lower the association between the two, the more intergenerational mobility exists, all else being equal. Structural mobility refers to overall shifts in the occupational mix, which affect the job prospects for individuals from different generations. Over time, as the economy has grown, the distribution of jobs has tilted toward the high-status end of the occupational spectrum, thereby improving intergenerational mobility for younger generations.
Intergenerational Mobility in the United States
In the sections that follow, we discuss research that has shed light on three important questions related to intergenerational mobility in the United States. First, how much intergenerational mobility is there? Second, how has this changed over time? Third, how does intergenerational mobility in the United States compare to other countries?
To summarize the evidence at the outset, there appears to be significant intergenerational mobility in the United States, although perhaps less than is sometimes believed. Origins significantly affect destinations. Specifically, adult sons and daughters are more likely to look like their parents in terms of occupation or income than one would predict on the basis of chance. Still, there is considerable mobility. Indeed, even when occupations or income categories are broadly defined, a majority of adult offspring occupy a different occupational or income category than their parents (even though the most likely single category is still that of their parents).
The causes of this intergenerational mobility have varied over time. Circulation mobility has increased in recent years, as the effect of family background on a child's eventual status has declined. At the same time, structural mobility has decreased, as the rate of economic growth has slowed, thereby reducing the likelihood that, on average, children will do better than their parents simply because of growth in the economy as a whole. On balance, these two sources of mobility have had offsetting effects. As a result, there has been little change in overall mobility in the past few decades.
Finally, international comparisons suggest that there are few systematic cross-national differences in intergenerational mobility, suggesting that, contrary to common belief, such mobility is not greater in the United States than it is in many other countries.
Occupational Mobility Across Generations
Intergenerational mobility can be measured by analyzing data across generations for either occupations or income. Both methods present some methodological difficulties, and the most accurate picture is likely obtained by combining findings from both types of analysis. The most popular method in sociology for measuring intergenerational mobility is to examine data on occupations across generations. In general, these studies take the form of analyzing the quantitative relationship between an individual's occupational origin and destination, after controlling for other factors that may affect an individual's occupational status. 11 Origin is frequently defined as the occupation of a parent at some earlier time (e.g., the father's occupation when the son or daughter was a young adult). More recent studies (i.e., most of those conducted since the early 1960s) use recursive models of socioeconomic status attainment, while earlier studies relied upon simpler analyses of occupational movement.
For purposes of analyzing occupational mobility, occupations are assigned by sociologists to categories in a hierarchical scale, with the number of categories varying across studies, ranging from five basic categories (upper white-collar, lower white-collar, upper manual, lower manual, and farm) up to much higher numbers. Blau and Duncan (1967) were among the first to use this method, and they ranked seventeen categories according to the proportion in the occupation with above (or below) average earnings and the proportion with a high school diploma. They report that these rankings are consistent with surveys of public perceptions of the status of various occupations. About five-sixths of the variation across occupations in status perception could be attributed to variation in income and education. Other studies that have ranked occupations in a hierarchy have used similar methods in defining occupational status. 12
Occupational data have a number of advantages over income data. First, an individual's occupation is generally more stable than his or her income. A doctor tends to remain a doctor, even if her income bounces around from year to year. Second, occupations are frequently better indicators of one's status in society than income alone, for they reflect additional factors that shape status most notably educational background. Thus, by using occupational data, researchers can distinguish a unionized crane operator who earns $60,000 from a lawyer who earns the same amount but enjoys higher social status. Such a distinction cannot be made using income data alone. Finally, relative to income data, occupational data across generations are more widely available and are available for a longer period of time and are probably more reliably reported.
At the same time, there are disadvantages to using occupational data. Most importantly, while occupational mobility can shed light on distinctions among individuals in social status, it can also obscure subtle but relevant differences, which are often related to income. For example, suppose a father and his son are both manual workers in manufacturing. It is possible that, at similar points in their careers, both enjoy a similar occupational status. But it is also possible that shifts in technology have raised the wages of the son relative to the father. In this case, occupational data may not tell the entire story, and income data would add significant new information.
Occupational Mobility Prior to the Middle of the Twentieth Century
Popular treatment of earlier eras of American history depict it as marked by tremendous opportunities for upward mobility typified best, perhaps, by the stories of Horatio Alger. 13 The existing evidence suggests, however, that intergenerational mobility during the nineteenth century was not as great as was commonly assumed.
There were no national surveys providing evidence on intergenerational mobility in the United States before the 1960s, so much of the evidence on mobility in this era has been drawn from local or regional studies by social historians. Analysis by Grusky (1986), using pooled data from several separate studies of mobility in selected cities, finds significant class-based intergenerational inequalities in the United States during the 19th century, with a particularly significant divide between the manual and nonmanual sectors in shaping individual life-chances. 14 Grusky also finds that the impact of class diminished over time, with rates of intergenerational mobility significantly higher in the 20th century than in the 19th century. 15 This increase in mobility was driven by increases over time in both circulation mobility and structural mobility.
Other studies focusing on the origins of the socioeconomic elite16 in America are generally consistent with this conclusion, finding that men who were successful came predominantly from wealthy origins. Pessen finds, for example, that of the "urban economic elite" in various northeastern cities during the Jacksonian era, only about 2 percent were born poor, and about 6 percent came from "middling" status. The remainder (over 90 percent) were born to well-to-do families. Similarly, studies of the 1870s (by Gregory and Neu) and of 1900 (by Miller) find that the "industrial elite" of those years were also born to great advantages and came from a remarkably narrow background marked by relatively high social standing. 17
Thernstrom (1970) examines intergenerational mobility for a sample of over 6,500 working-age male residents of Boston between 1880 and 1963,18 and finds that approximately 40 percent of the sons of working-class fathers had moved up into middle-class jobs by the ends of their careers. Over the 80 or so years that he analyzed, rates of intergenerational mobility were remarkably constant. (See Table 1 for a more detailed description of the methodology of this study and the other studies of intergenerational mobility discussed below.)
Other evidence suggests that overall intergenerational mobility for males may have increased somewhat during the first half of this century. 19 In a classic study, Lipset and Bendix (1959) analyzed 1949 data on the job histories and socioeconomic origins of a sample of over 600 male workers in Oakland, California. Examining the broad occupational categories in which individuals had worked at some time in their careers, they found significant intergenerational mobility, as well as evidence that many individuals worked in a wide variety of occupations in a lifetime (suggesting considerable mobility within a given generation).
They found significant discrepancies between the careers of sons growing up in families in which fathers worked in a high status occupation and those growing up in a working class family. For example, 69 percent of the sons of professional or semi-professional fathers spent at least some time in a high status job, and 45 percent spent some time in a working class job. These numbers compare to 42 percent of the sons of semi-skilled or unskilled fathers spending some time in a high status job and 80 percent spending some time in a low status job. Still, after observing the differences, Lipset and Bendix suggest that "it is perhaps even more significant that [the differences] are not greater" (1959, 185).
The higher mobility observed during the first half of this century appears to have been driven primarily by the growth in structural mobility. The distribution of occupations changed dramatically (see Chart 1) throughout much of this century, as farming and other forms of self-employment declined while production and, later, white-collar employment increased. As a result of these changes, successive generations faced very different labor markets (see Chart 2). Overall, workers were redistributed into occupations that were, on average, better than those of their parents (Hauser et al. 1975). This was not, however, a result of a decrease in the dependence of occupational destinations on origins; rather, it was simply a result of there being a larger number of better jobs available.
Overall, the existing evidence suggests that mobility was likely not as great as suggested by popular literature and the writings of Tocqueville on the openness of American society. Most of the rich during earlier periods were apparently born rich. It also appears that class was more important during the nineteenth century than it is today.
Occupational Mobility in the Second Half of the Twentieth Century
The first comprehensive analysis of occupational mobility was made possible by the 1962 and 1973 Occupational Changes in a Generation (OCG) surveys conducted by the U.S. Bureau of the Census. The 1962 OCG includes data for over 20,000 men between the ages of 20 and 64, and the 1973 OCG includes data for over 33,000 men between the same ages. These much-improved data on occupational destinations and origins made possible more sophisticated analysis of trends in intergenerational mobility.
Analysis of the 1962 data found that education represented the primary vehicle through which a father's status (education and occupation) was transmitted to a son. Education as a transmitter of fathers' status had a greater impact than the independent effect of father's status. Still, the most important overall factor shaping sons' status was the effect of education, independent of fathers' background (Blau and Duncan 1967).
Featherman and Hauser (1978) compared the 1962 and 1973 data in an effort to determine change over time. They found that the percentage of variation in son's occupational achievement that could be explained by education (independent of father's background) increased from 19 to 22 percent. The percentage of variation directly attributable to the independent effect of father's background fell from 11 percent to 7 percent, while the overlapping influence of the two (i.e., the effect of father's background, as transmitted through education) remained constant at 14 percent. In both years, a similar majority of the variation in occupational achievement is unexplained by those factors. 20 Thus, the overall effects of origins on individual destinations declined slightly, and circulation mobility increased. 21
In more recent years, circulation mobility has continued to increase, as the relationship between individual origins and destinations has continued to weaken. At the same time, however, the earlier growth in structural mobility has slowed considerably in the last two decades, largely offsetting the positive effect of circulation mobility on overall intergenerational mobility.
Some of the most important evidence on these issues is provided by Hout (1988). He examines the association between origins and destinations by using data from the General Social Survey from the National Opinion Research Center. This data set includes data on over 9,000 men and women between the ages of 25 and 64 who were in the civilian labor force between 1972 and 1985. Hout assigns occupations to 14 categories based primarily on the 17-occupation structure that was first used by Blau and Duncan (1967).
He finds that the association between origins and destinations decreased by one-third between 1972 and 1985. He attributes this increase in circulation mobility to the rising percentage of workers who have college degrees (see Chart 3). He finds that for such workers, there is no relationship between origins and destinations, so the increasing prevalence of college graduates in the labor force increases the overall rate of intergenerational mobility, all else being equal.
The relationship between origins and destinations for non-college graduates, however, has remained generally constant. Chart 4 depicts graphically the changing relationship between fathers' and sons' statuses. As the average association between the two declines, the lines become more horizontal. For college graduates (i.e., those for whom there is no destination-origin relationship), the line is horizontal.
Hout also finds that overall distributions of origins and destinations are growing more similar. This suggests that the historical increase in structural mobility discussed above is slowly declining. Hout finds that this decline almost exactly offset the increase in circulation mobility, resulting in overall intergenerational mobility rates being unchanged from the early 1970s to the middle of the 1980s. Supplemental analysis by the current authors of the data used by Hout (reported in Table 2) shows the continuing effect of origins on destinations (see Chart 5).
This supplemental analysis also provides information on the changing reasons for intergenerational mobility. In the most recent period covered by Hout's study (1982-85), approximately 65 percent of intergenerational mobility can be attributed to circulation mobility, compared with only 42 percent during the earlier period, 1972-75 (see Chart 6). The remainder in each period can be attributed to productivity growth and related changes in the structure of the economy. 22
The findings of Grusky and DiPrete (1990a) support the conclusions of Hout. Grusky and DiPrete also analyze data from the General Social Survey, using data from over 8,000 respondents between the ages of 25 and 64 for the years 1972 and 1987. 23 They find that the effects of social background (as well as race) are declining over time, and also detect the slowing in the rate at which intergenerational mobility has been occurring as the result of economic growth.
Similarly, Biblarz, Bengston, and Bucur (1996), using a different data set, find that the association between parents' socioeconomic status and that of their children has weakened over time. Using a sample of over 2,000 individuals from the Longitudinal Survey of Generations, they examine changes in intergenerational mobility across a number of generations. Because the survey was originally based on the members of a Los Angeles-area health care plan, the results are not as generally applicable as the results from studies that are based on more broadly representative data. They are interesting, however, because they cover three generations, the earliest of which reached adulthood in 1921.
The authors examine occupational mobility across these three generations. Their data support the conclusion that circulation mobility has increased across generations. Like the studies discussed above, Biblarz, et al. also find increases across generations in the overall occupational positions achieved by each generation, reflecting the results of economic growth. The relative similarity of the occupational distributions of the two most recent generations, however, suggests that structural mobility has decreased, partially offsetting the effects of increased circulation mobility. The authors conclude that newer generations may face increased difficulty in experiencing the upward mobility enjoyed by previous generations.
Recent analysis by Hauser, et al. (1996), using data from the GSS, OCG, and also the Survey of Income and Program Participation (SIPP), yields results consistent with the studies discussed above. The authors find that circulation mobility has increased at the same time as overall upward intergenerational mobility has been declining. They also suggest that analysis of the extent to which siblings end up in similar positions would be a better method for analyzing the extent of opportunity in society. Carrying out perhaps the first large-scale analysis of this sort, they conclude that correlations among siblings in occupational status (as well as educational attainment) are surprisingly low, suggesting a significant degree of overall opportunity.
Thus, the overall evidence on trends in intergenerational mobility, as measured by occupations, appears to be clear. Since at least the mid-1970s, there has been an increase in circulation mobility, meaning that the links between individuals' destinations and origins are eroding. At the same time, however, the rapid change in the occupational structure bringing with it an increasing number of high-status jobs has been slowing. These trends appear to have substantially offset one another, resulting in little overall change in the rate of intergenerational occupational mobility for the population as a whole.
The findings have different implications, however, for various subgroups. Those from modest backgrounds are as likely to move up as they were previously because the effects of slower economic growth have been offset by greater circulation mobility. But those from more privileged backgrounds are adversely affected by both the slowdown in economic growth and the increase in circulation mobility. For this group, the increase in circulation mobility means a higher probability of moving down the ladder.
Intergenerational Income Mobility
Researchers also measure intergenerational mobility by looking at the relationship between the incomes of parents and the incomes of their children at corresponding points in their careers. 24 Indeed, as Solon (1992, 393) notes, "given the widespread concern about intergenerational mobility, it is astonishing how few efforts have been made to measure the simple intergenerational correlation of income in the United States."
Solon (1989, 1992) uses a sample of 348 father-son pairs from the Panel Study of Income Dynamics (PSID). He averages fathers' incomes over the five years from 1967 to 1971 (when the average age of the fathers was 42 to 46) and sons' incomes in 1984 (when the average age of the sons was 30). 25 He finds that the intergenerational income correlation in the United States is quite significant approximately 0.4, indicating that approximately 16 percent of overall variation in sons' incomes may be attributed to fathers' incomes. 26 The correlation between total family income in the earlier period and sons' income in 1984 was 0.5, even higher than the correlation between the individual incomes of fathers and sons.
The effects of such a correlation on individual opportunity are significant. Assuming a normal distribution of long-term income, the correlation of 0.4 would imply that a son whose father's income was at the 25th percentile would have a 26 percent chance of being in the bottom quintile, a 39 percent chance of rising above the median, and a 12 percent chance of moving up to the highest income quintile (see Chart 7). A son whose father's income was at the 95th percentile would have a 5 percent chance of being in the bottom quintile, a 76 percent chance of being above the median, and a 42 percent chance of being in the top quintile. 27 Thus, Solon's findings indicate that there remains a substantial component of income immobility in the United States.
Zimmerman (1992) uses data on over 800 father-son pairs from the National Longitudinal Survey, with data from 1966 through 1981. He compares data for fathers in the 1960s (when they were around age 50) and sons in the 1980s (when they were in their early to mid-30s), after making an effort to control for the different ages of the two generations. Also making adjustments for the purposes of averaging fathers' income, he finds results that are quite consistent with those of Solon a correlation of approximately 0.4 between the incomes of fathers and sons. He finds similar correlations for the Blau-Duncan measure of occupational status.
Thus, evidence on intergenerational mobility, as measured by income, indicates that the links between the income of parents and the later incomes of their adult children remain significant, although there is no information on whether this relationship (like occupational mobility) has changed over time.
The observed correlation of 0.4 implies that 16 percent of variation in children's income is associated with variations in parental income. How should this finding be interpreted? Is it higher than we would expect in a society marked by significant equality of opportunity? We return at this point to our earlier estimate of how much opportunity we would expect to observe in a society in which only the inheritance of cognitive ability constrains individual opportunity. Recall that in such a society, a maximum of 4 percent of the variation in children's incomes might be associated with parental characteristics. Solon's and Zimmerman's findings indicate that the actual association between parents' and children's incomes is much higher. This indicates that there are significant familial factors beyond the simple inheritance of ability that fundamentally shape economic outcomes. Thus, society's playing field is tilted significantly toward individuals born and raised in families that can, in some way, provide them with advantages not available to comparably-talented individuals from lower-income families.
How does intergenerational mobility in the United States compare to that in other nations? A number of studies have attempted to answer this question, and generally find that there are fewer cross-national differences than one might expect. We briefly discuss some of this analysis below.
As early as 1959, Lipset and Bendix found that there was relatively little difference in rates of social mobility (as measured by movement between manual and nonmanual occupations) across 9 industrialized countries. 28 Although they observed some differences in mobility across countries, they concluded that there was no evidence to support the popular view that the United States was the land of opportunity. They argued that if the United States is unique in any way, it is in the way in which the American public perceives and evaluates mobility. Because opportunity represents an integral component of the American credo, Americans are more inclined to take a positive view of the mixed evidence on how much mobility is present in society, seeing it as a reflection of the openness and equality of opportunity in American life, and not as proof of the persistence of inequalities.
Others, however, argued that there are certain distinctive qualities to mobility in America. Blau and Duncan (1967) concur with the general observation that mobility is not significantly different in the United States than in other countries, but suggest that opportunities for substantial upward mobility in the United States (i.e., from the lowest social strata into a top stratum) are greater than in other countries. Based on comparisons of his analysis of working-class mobility in Boston and studies of European cities, Thernstrom (1974) argues that Americans are more likely to be able to move up from working-class origins.
More recent analyses are generally consistent with the Lipset and Bendix thesis. Erikson and Goldthorpe (1993) and Ganzeboom, et al. (1991) review cross-national data on intergenerational mobility and find that, although there are some differences in cross-national mobility rates, there are few broad conclusions that can be drawn regarding explanatory factors. 29 Erikson and Goldthorpe find national mobility rates are remarkably stable over time, with only a few countries displaying any evidence of increasing mobility. Ganzeboom, et al., on the other hand, find that intergenerational mobility has increased slowly and systematically in most countries since World War II.
Kerckhoff, et al (1985) compare intergenerational mobility rates in two countries: the United States and Great Britain. Using data from the 1970s, they find significant shifts in both countries in occupational distribution between the fathers' and sons' generations. Particularly notable were the shifts in the United States away from farming and toward white-collar occupations. They find that more individuals in the United States move up in the occupational hierarchy than move up in Britain, but ultimately conclude that this difference is explained by differences in economic growth, not by differences in the social structures that determine opportunity and achievement in each country. Thus, if occupational trends were the same in the two countries, measures of intergenerational mobility would also have been much the same. 30
Overall, the evidence suggests that the playing field is becoming more level in the United States. Socioeconomic origins today are less important than they used to be. Further, such origins have little or even no impact for individuals with a college degree, and the ranks of such individuals continue to increase. This growth in access to higher education represents an important vehicle for expanding opportunity. Still, family background continues to matter. While the playing field may be becoming more level, family factors still significantly shape the economic outcomes of children.
As Blau and Duncan (1967, 430) note, we now live in a society in which "the achieved status of a man, what he has accomplished in terms of some objective criteria, becomes more important than his ascribed status, who he is in the sense of what family he comes from. This does not mean that family background no longer influences careers. What it does imply is that superior status cannot any more be directly inherited but must be legitimated by actual achievements that are socially acknowledged. Education assumes increasing significance for social status in general and for the transmission of social standing from fathers to sons in particular."
How does the evidence of growing opportunity fit together with the evidence that young adults today have lower incomes than their parents did at a similar age? First, it should be recognized that an increase in intergenerational mobility in the form of greater circulation mobility would not be expected to increase the incomes of younger generations. Rather, it merely frees individuals to succeed or fail on their own, independent of the influence of their socioeconomic background. Only an increase in structural mobility (reflecting an increase in productivity growth) would be expected to have a net positive effect on the average incomes of newer generations.
Since evidence indicates that structural mobility growth has decelerated in recent years, it would be expected that income growth for younger generations would slow down as well. For the youngest cohorts currently in the labor force, this trend has been exacerbated by other factors, including increases in the wage premium for work experience. Further, young men in particular have seen their positions eroded in comparison to young women.
Thus, the widespread perception that the youngest generation of workers men, in particular is not doing as well economically as their parents did is partly a reflection of the more static occupational structure that has paralleled the slowdown in economic growth. At the same time, children today are increasingly free to move beyond their roots equality of opportunity has increased. Yet there are fewer places for them to go that represent an improvement over the positions of their parents.
How any particular individual is affected by these trends will vary. The most negative effects are likely to be felt by individuals from more privileged backgrounds. For this group, the increase in circulation mobility now makes it more likely that they will move down the ladder (the only direction in which they can move), while slower growth has constrained their opportunities to move further up. In contrast, those who grew up in more modest circumstances are more likely to be able to move up, although slower growth has created fewer places for them to go. Overall, the slowdown in economic growth has created a situation in which fewer individuals are now able to surpass their parents' standard of living, despite the decreased importance of family background on ultimate status.
Behrman, Jere R., and Paul Taubman. 1990. "The Intergenerational Correlation Between Children's Adult Earnings and Their Parents' Income: Results from the Michigan Panel Survey of Income Dynamics." Review of Income and Wealth 36 (June 1990): 115-127.
Biblarz, Timothy J., Vern L. Bengtson, and Alexander Bucur. 1996. "Social Mobility Across Three Generations." Journal of Marriage and the Family 58 (February 1996): 188-200.
Blackburn, McKinley L., and David Neumark. 1993. "Omitted-Ability Bias and the Increase in the Return to Schooling." Journal of Labor Economics 11: 521-544.
Blau, Peter M., and Otis D. Duncan. 1967. The American Occupational Structure. New York: Wiley.
Bode, Carl. 1985. "Introduction." In Horatio Alger, Ragged Dick and Struggling Upward. New York: Penguin.
Devlin, Bernie, Stephen Fienberg, Daniel Resnick, and Kathryn Roeder. 1995. "Galton Redux: Eugenics, Intelligence, Race, and Society: A Review of The Bell Curve: Intelligence and Class Structure in American Life." Journal of the American Statistical Association (December): 1483-1488.
Dickens, William, Thomas Kane, and Charles Schultze. 1997. "Does The Bell Curve Ring True? A Reconsideration." Unpublished manuscript.
Duncan, Otis Dudley. 1975. Introduction to Structural Equation Models. New York: Academic Press.
Duncan, Otis Dudley, David L. Featherman, and Beverly Duncan. 1972. Socioeconomic Background and Achievement. New York: Seminar Press.
Erikson, Robert, and John H. Goldthorpe. 1993. The Constant Flux: A Study of Class Mobility in Industrial Societies. New York: Oxford University Press.
Featherman, David, and Robert Hauser. 1978. Opportunity and Change. New York: Academic Press.
Flynn, James R. 1996. "IQ Trends Over Time: Intelligence, Race, and Meritocracy." Mimeograph. University of Otago, Dunedin, New Zealand.
Ganzeboom, Harry B.G., Donald J. Treiman, and Wout C. Ultee. 1991. "Comparative Intergenerational Stratification Research: Three Generations and Beyond." Annual Review of Sociology 17: 277-302.
Goldberger, Arthur S., and Charles F. Manski. 1995. "Review Article: The Bell Curve by Herrnstein and Murray." Journal of Economic Literature 33 (June 1995): 762-776.
Grusky, David B. 1986. "American Social Mobility in the 19th and 20th Centuries." University of Wisconsin Center for Demography and Ecology. Working Paper No. 86-28.
Grusky, David B., and Thomas A. DiPrete. 1990a. "Recent Trends in the Process of Stratification." Demography 27 (November 1990): 617-637.
Grusky, David B., and Thomas A. DiPrete. 1990b. "Structure and Trend in the Process of Stratification for American Men and Women." American Journal of Sociology 96 (July 1990): 107-143.
Grusky, David B., and Ivan K. Fukumoto. 1989. "Social History Update: A Sociological Approach to Historical Social Mobility." Journal of Social History 23: 221-232.
Hauser, Robert M., Peter J. Dickinson, Harry P. Travis, and John M. Koffel. 1975. "Temporal Changes in Occupational Mobility: Evidence for Men in the United States." American Sociological Review 40 (June 1975): 279-297.
Hauser, Robert M., and John Robert Warren. 1996. "Socioeconomic Indexes for Occupations: A Review, Update, and Critique." University of Wisconsin Center for Demography and Ecology. Working Paper No. 96-01.
Hauser, Robert M., John Robert Warren, Min-Hsiung Huang, and Wendy Y. Carter. 1996. "Occupational Status, Education, and Social Mobility in the Meritocracy." University of Wisconsin Center for Demography and Ecology. Working Paper No. 96-18.
Haveman, Robert. 1987. Poverty Policy and Poverty Research: The Great Society and the Social Sciences. University of Wisconsin Press.
Herrnstein, Richard, and Charles Murray. 1994. The Bell Curve: Intelligence and Class Structure in American Life. Free Press.
Hout, Michael. 1988. "More Universalism, Less Structural Mobility: The American Occupational Structure in the 1980s." American Journal of Sociology 93 (May 1988): 1358-1400.
Hout, Michael, and Robert M. Hauser. 1992. "Symmetry and Hierarchy in Social Mobility: A Methodological Analysis of the CASMIN Model of Class Mobility." European Sociological Review 8: 239-266.
Jencks, Christopher. 1972. Inequality. New York: Basic Books.
Kerckhoff, Alan, Richard Campbell, and Idee Winfield-Laird. 1985. "Social Mobility in Great Britain and the United States." American Journal of Sociology 91: 281-307.
Lipset, Seymour Martin, and Reinhold Bendix. 1959. Social Mobility in Industrial Society. Berkeley: University of California Press.
Mare., Robert D. 1997. "Differential Fertility, Intergenerational Educational Mobility, and Racial Inequality." University of Wisconsin Center for Demography and Ecology. Working Paper No. 97-03.
McMurrer, Daniel P., and Isabel V. Sawhill. 1996. "Economic Mobility in the United States." Washington, DC: The Urban Institute.
Pessen, Edward (ed.) 1974. Three Centuries of Social Mobility in America. Lexington, MA: Heath and Company.
Richmond, Samuel B. 1964. Statistical Analysis. Second Edition. New York: Ronald Press.
Sawhill, Isabel V., and Daniel P. McMurrer. 1996. "Declining Economic Opportunity in America." The Future of the Public Sector No. 1. Washington, DC: The Urban Institute.
Sewell, William H., and Robert M. Hauser. 1975. Education, Occupation, and Earnings: Achievement in the Early Career. New York: Academic Press.
Solon, Gary. 1989. "Intergenerational Income Mobility in the United States." University of Wisconsin-Madison IRP Discussion Paper #894-89.
Solon, Gary. 1992. "Intergenerational Income Mobility in the United States." American Economic Review 82 (June 1992): 393-408.
Thernstrom, Stephan. 1970. "Working Class Social Mobility in Industrial America." In Melvin Richter (ed.), Essays in Theory and History. Cambridge, MA: Harvard University Press, pp. 221-238.
Thernstrom, Stephan. 1974. "Socialism and Social Mobility." In John Laslett and Seymour Martin Lipset (eds.), Failure of a Dream? Essays in the History of American Socialism. New York: Anchor, pp. 509-551.
Zimmerman, David J. "Regression Toward Mediocrity in Economic Stature." American Economic Review 82 (June 1992): 409-429.
Tables and Charts
Table 1. Comparison of Selected Studies of Intergenerational Mobility
||BIRTH COHORTS OF OFFSPRING IN SAMPLE
|Lipset and Bendix (1959)
||644 male "principal wage earners" in Oakland, California
||U. of California Institute of Industrial Relations data (1949)
||Not clear. If ages 20-65, then would have been born 1884-1929
||"Principal" occupation of father is compared to son's occupations over entire career.
||Significant differences in careers of sons, depending on whether middle-class or working-class. Educational attainment as major determinant of career patterns.|
||6500 working age males, Boston
||City census files for males who also appeared in files 10 years later (approximately 50 percent of all males in original samples)
||Not clear. Working age males between 1880-1963. If ages 20-65, then would have been born in 1815-1943
||Intergenerational occupational mobility relatively constant over years examined. Approximately 40 percent of sons of working class move into middle-class jobs by end of career.|
|Blau and Duncan (1967)
||20,700 men between ages 20 and 64 in 1962
||Occupational Change in a Generation (1962 only)
||Born between 1898-1942
||Father's socioeconomic characteristics (education and occupation when respondent was 16) are compared to respondent's current job and first job.
||Son's education as most important channel through which father's education and occupation influenced son's status. This is more important than father's status overall, but less important than independent effect of education.|
|Featherman and Hauser (1978)
||total of 54,300 men in two surveys (20,700 in 1962, and 33,600 in 1973) between ages 25 and 64
||Occupational Change in a Generation (1962 and 1973)
||First group born between 1898-1937; Second group born between 1909-48
||Father's socioeconomic characteristics (education and occupation when respondent was 16) are compared to respondent's current job and first job. Results are compared for the two (overlapping) samples.
||Between 1962 and 1973, the independent effect of education on son's occupation increased, and independent effect of father's background decreased.|
||9,227 civilian workers between ages 25 and 64 in years of survey
||3 data periods (1972-75, 1976-80, 1982-85) from General Social Survey
||Born between 1908-60
||Father's occupation (as reported by respondent) is compared with current occupation of respondent.
||Between 1972 and 1985, circulation mobility increased, as the association between origins and destinations fell by one-third. This is due in large part to the increase in the percentage of college graduates in the labor force. A decrease in structural mobility offset the effects of increased circulation mobility.|
|Grusky and DiPrete (1990a)
||8,127 civilian workers between ages 25 and 64 in years of survey
||Annual data from 1972-87 from General Social Survey
||Born between 1908-62
||Father's occupation when respondent was "growing up" is compared with "normally held" occupation of respondent.
||Effects of socioeconomic background decreased between 1972 and 1987.|
|Biblarz, Bengston, and Bucur (1996)
||1,142 individuals in 3 generations
||Longitudinal Survey of Generations
||Eighty percent of each generation born between the following years:
|Father's occupation when respondent was teenager is compared to respondent's own occupation when respondent's children are teenagers. (For G3s, however, current occupation in 1985 was used, when many offspring of G3s would be expected to be teens.)
||Decrease in overall association of socioeconomic status between parents and children across generations examined (i.e., more mobility for G3 than G1). Circulation mobility has increased, structural mobility has decreased.|
||348 father-son pairs
||Panel Study of Income Dynamics
||Oldest son in each sample family who was born between 1951-59
||Son's earnings in 1984 (when average age of sons was 30) are compared to average of father's earnings in 1967-71 (when average age of fathers was 42-46)
||Correlation of income across generations is 0.4 to 0.5.|
||876 father-son pairs
||National Longitudinal Survey
||Oldest son in each sample family average year of birth is 1947
||Various comparisons controlling for life-cycle effects; closest ages are fathers in 1965 (average age 50) and sons in 1981 (average age 34)
||Correlation of earnings across generations is at least .40; comparable correlation for occupational status|
Occupational Transitions of Males, 1972-75 and 1982-85
Source: Authors' calculations based on Hout (1988) data.
1. For example, men born in the 1940s, who were between the ages of 25 and 34 in 1974, earned a median income of about $30,000 that year (in 1994 dollars). Men born two decades later in the 1960s, who were 25 to 34 years old in 1994 earned less than $23,000 at a comparable age, representing a precipitous drop in earnings (Sawhill and McMurrer 1996). Note that this is an intercohort comparison and not, strictly speaking, a comparison between fathers and sons.
2. It is possible that there are other macroeconomic factors independent of economic growth that affect intergenerational mobility independent of economic growth. Sociologists make frequent reference to the evolution or maturation of the economy (i.e., the shift from a manufacturing economy to a post-industrial economy), often with the implicit suggestion that this economic evolution may have an independent effect on mobility. Such factors are inextricably linked to the growth rate of the economy, however, and it is not clear whether they have any quantifiable effects that are independent of the rate of growth. In this paper, therefore, we will view all macroeconomic influences on intergenerational mobility as reflections of economic growth.
3. Overall rates of intergenerational mobility can also be affected by different rates of fertility among different socioeconomic groups, which could create broad flows of individuals up (or down) the economic spectrum across generations.
4. See Behrman and Taubman (1990, 117-118).
5. Such a society would not necessarily be the most mobile. A society that randomly assigns economic positions will have the most mobility (i.e., the weakest relationship between origins and destinations). Such a society, however, would be neither appealing nor viable, for it would be highly inefficient. Thus, perfect social mobility would have its economic costs, just as inadequate social mobility does.
6. For example, if fertility and income are negatively correlated, an upward mobility trend could still prevail even in the absence of any economic growth. This would occur as vacancies in occupations would necessarily be filled by the (more numerous) individuals from disadvantaged backgrounds, many of whom would be moving up in the distribution.
It is interesting to note that the differential rates of fertility among various demographic groups have been viewed by some as an important means of sustaining upward mobility in society, while they have been cited by others as likely to bring about "dysgenesis," downward pressure on the overall level of intellectual ability within society. For various perspectives, see analysis and discussion by Mare 1997, Flynn 1996, and Herrnstein and Murray 1994 (esp. pp. 341-368).
7. For a full discussion of public opinion on this issue, see Herbert McClosky and John Zaller, The American Ethos: Public Attitudes Toward Capitalism and Democracy (Cambridge: Harvard University Press, 1984), and Seymour Martin Lipset, American Exceptionalism (New York: W.W. Norton, 1996).
8. This estimate was derived, as discussed below, from estimates of (1) the correlation between parents' and children's cognitive ability, and (2) the correlation between cognitive ability and earnings.
The observed correlation between the Intelligence Quotients (IQs) of parents and children is believed to be somewhere between 0.34 and 0.50, implying that 12 to 25 percent of variation in individual IQs is associated with parental IQs (the percentage of total variation in one variable that is associated with another variable is the square of the correlation between the two variables). Some of this is due to the influence of genetics, while some is due to environment. These figures are derived from estimates by Goldberger and Manski (1995) and Devlin, et al. (1995).
Goldberger and Manski estimate that the correlation between parents' and children's IQs is 0.4 to 0.5. Examining the evidence from the perspective of quantitative genetics, Devlin and co-authors estimate that the correlation between the two is between 0.34 and 0.45, depending on what type of correlation is being examined (the lower number represents a closer estimate of the extent of variation that is associated exclusively with genetic differences).
Observed relationships between earnings and various measures of cognitive ability or intelligence generally imply correlations of 0.28 to 0.40, suggesting that as much as 8 to 16 percent of the variation in wages is associated with variation in intelligence or ability. These figures are derived from estimates by Dickens, Kane, and Schultze (1997) and Blackburn and Neumark (1993).
Dickens and co-authors find that the scores on the Armed Forces Qualifying Test (controversially used by Herrnstein and Murray as a measure of intelligence) account for 12 percent of variation in annual earnings and 16 percent of variation in family income. The estimates do not control for other factors (except for age), which would cause the estimated associations to decline. Analysis by Blackburn and Neumark (1993) found that the effects of test scores accounted for 8 to 11 percent of the variation in wage rates in 1987. Earlier analysis by Jencks (1972) using 1962 data were consistent with the above results: he found that test scores accounted for 12 percent of the variation in family incomes. The correlation estimates reported in the text are those that are consistent with the amount of explained variation, as discussed above.
The two sets of estimates (parent-child IQ and ability-wages) can be combined to yield a very rough upper-bound estimate of the extent of variation in earnings that could be associated with the genetic transmission of cognitive ability. By multiplying the two respective correlations together (see Duncan 1975, 9-13, for more discussion of this technique), we arrive at an upper-bound estimate of 0.20 for the correlation between parents' IQ and children's earnings, which implies that not more than 4 percent of total variation in earnings can be associated with genetic transmission.
This estimate assumes that tests such as the AFQT are able to capture the genetically-transmitted "ability" that affects earnings. If there are other attributes (e.g., health or other types of ability) with a genetic component that can affect one's earnings but are not captured in such tests, then this estimate could understate the maximum association between parents' characteristics and children's income. On the other hand, it is generally recognized that AFQT scores are not a pure measure of inherent ability or IQ, and that even the latter has an environmental component that may be influenced by schooling or other external factors.
9. The (hypothetical) society in which the transmission of some component of cognitive ability from parents to children was the only form of intergenerational association would thus be the society that provides the most level playing field possible (assuming that no efforts are made to offset genetic differences). Thus, our estimate of the proportion of variation associated with genetic factors in such a society represents a baseline against which we can evaluate the implications of the degree of intergenerational mobility present in the United States. We will return to this estimate later in the paper.
10. This question is explored in additional detail later in this paper.
11. Until recent years, data were available only for males, so most of these studies analyzed the relationship between fathers' and sons' occupations.
In addition, note that comparisons of origins and destinations at a single moment in time are unable to capture the full extent of intergenerational mobility for an individual who either has a highly mobile career or has parents who had mobile careers. See Lipset and Bendix (1959).
12. For a discussion the creation and use of occupational indices, see Hauser and Warren (1996).
13. Alger wrote over one hundred novels. Almost all of these tell stories of individuals moving upward from rags to respectability (but not necessarily riches) in an America in which good fortune is the reward for the virtuous and hard-working young man. For additional discussion, see Carl Bode, "Introduction," in Horatio Alger, Ragged Dick and Struggling Upward, New York: Penguin Books (1985 edition).
14. See Grusky and Fukumoto (1989).
15. For 20th century data, Grusky uses data from the 1962 and 1973 Occupational Changes in a Generation surveys. These surveys are discussed in more detail later in this paper.
16. Many studies of early mobility focused on the origins of the wealthy, because records (including tax records) are much more readily available for these individuals than for any others. The data are therefore necessarily less representative of the population as a whole than are more recent data. It is entirely possible that there were other trends in mobility that were simply not picked up due to the limited nature of the available data.
17. These three studies (Pessen, Gregory and Neu, and Miller) are collected in a work edited by Pessen (1974).
18. Because this sample includes only those individuals who actually settled down long enough to have their careers measured, Thernstrom suggests that the mobility rates he finds may be somewhat inflated.
19. For additional discussion of earlier studies, see Hout (1988) and Pessen (1974).
20. For a more detailed discussion of the methodologies and results of this early "status attainment" research, see Haveman (1987, 108-116).
21. Because these subtle changes were observed over a period of only eleven years, it is unwise to draw broad conclusions from these results. Nevertheless, when they are combined with later studies (some of which also analyze only a relatively short time period), a consistent picture does emerge of the changes in mobility that have occurred over the last few decades.
22. These numbers were derived from the following calculations: in both time periods, 65 percent of the labor force is mobile (either up or down) across five major occupational categories derived from those used by Hout. The number of individuals who are upwardly mobile would have to be equal to the number who are downwardly mobile if there were no change in the occupational distribution in a given period. Thus, the difference in each period between those who are upwardly and downwardly mobile can be read as the segment of mobility that is due to economic growth, while the remainder is due to circulation mobility.
23. Unlike Hout, who had to pool annual data into three broader periods, the model employed by Grusky and DiPrete is simple enough that they are able to examine year-to-year changes.
24. One approach is to determine the relationship between a father's income in a given year and his son's income in another year. Unfortunately, the resulting measure of intergenerational mobility is probably misleading, since individual incomes rise and fall significantly from year to year (see Daniel P. McMurrer and Isabel V. Sawhill, "Economic Mobility in the United States," Urban Institute, 1996, No. 6722). As a result, a person's income in a given year is likely to be a poor approximation of his or her long-term economic position. Thus, measures of intergenerational mobility that are based on annual income figures may be unreliable. Our discussion below focuses on those studies that use more than one year in calculating income.
25. By averaging fathers' incomes over five years, Solon hopes to develop a more accurate indicator of their actual long-term economic positions.
26. This estimate is substantially higher than was found in earlier research. Solon argues that earlier attempts (e.g., Behrman and Taubman 1985 and Sewell and Hauser 1975) to analyze intergenerational income mobility were flawed for two methodological reasons. First, they used either single-year income measures or measures averaged over a few years, thereby biasing correlation estimates downward (because the single-year measures, as noted above, are noisy). Second, they used data samples that were unrepresentatively homogeneous, resulting in what Solon suggests are downward biases in the correlation coefficient.
27. Solon does finds some tentative evidence that individuals who grew up in the bottom of the income distribution are more likely to move up in the distribution than individuals who grew up at the top are likely to move down. If this were the case, the probabilities cited above would be slightly different (specifically, the son growing up at the bottom of the distribution would be slightly less likely to remain near the bottom).
28. They did suggest that there is more mobility into the ranks of the elite in the United States.
29. For a methodological critique of some of this analysis, see Hout and Hauser (1992).
30. Citing the evidence of changing intergenerational dynamics in the United States, Hout (1988) expresses concern about using "snapshot" pictures of national mobility.