Last Updated on December 28, 2021

In this post, I explore racial disparities in intergenerational mobility, i.e. racial disparities in offspring outcomes after controlling for parental achievement on the same outcome. The primary focus is on black-white disparities in income mobility, i.e. the finding that black children earn lower incomes than white children with similar parental incomes. Other racial groups and socioeconomic outcomes will be considered when data is available. I start by documenting racial disparities in various socioeconomic outcomes, such as income, educational attainment, and wealth. I also show that there are also large racial disparities in mobility for each of these outcomes. Next, I document some of the patterns of the racial mobility gaps in more detail, by showing the history of the gaps, gender differences in the gaps, and comparisons with racial groups other than blacks and whites. Following that, I explain why differences in income mobility are pivotal to explaining persistent income gaps between blacks and whites. I then consider a number of different factors that might explain black-white gaps in income mobility. Finally, I end by considering what I take to be important implications of these findings.

Summary

Because the post is rather long, I’ll summarize the main points of each section here:

Patterns of racial mobility gaps

• The black-white gap in individual income mobility among males has remained fairly constant throughout the 20th century and has reduced only slightly within the past few decades.
• Controlling for parental household income, the black-white gap in individual income ranks is just one-third as large as the gap in household income ranks, likely because blacks are far less likely to marry even when conditioning on parental income.
• When considering more racial groups, Asian Americans have the greatest mobility, followed by whites, then Hispanics, with blacks and American Indians having the lowest levels of mobility.
• Many black-white mobility disparities are rather large for men but small (and sometimes non-existent or reversed) for women.

The importance of the black-white income mobility gap

• The income mobility gap, rather than income gaps of previous generations, is the primary proximate cause of the raw income gap today.
• The current income gap is in a steady state. If the current gap in mobility does not change, then the income gap is not expected to reduce any further.
• If the racial gap in mobility was eliminated, the vast majority of the income gap would be eliminated within just two generations.

Findings from my original analysis of black-white mobility gaps

• The current race-specific income mobility rates suggest that contemporary income gaps can emerge within just 2-3 generations even if blacks and whites start with perfectly equal income distributions. Similarly, contemporary income gaps can vanish within just 2-3 generations if blacks had similar rates of income mobility as whites.
• Using the same methodology as other papers in the mobility literature, I estimate the steady states of black and white wealth distributions. Further, I show that, like income, current race-specific wealth mobility rates suggest that contemporary wealth gaps can emerge within just 2-3 generations even if blacks and whites start with perfectly equal income distributions. Similarly, contemporary wealth gaps can vanish within just 2-3 generations if blacks had similar rates of wealth mobility as whites.

Explaining the black-white income mobility gap

• Family-level factors: parental education and parental wealth explain very little (<20%) of the mobility gap. Family structure (i.e. marital status, single mother household) may explain a substantial portion of the upward mobility gap, although only if using measures of family structure throughout one’s entire childhood rather than a particular snapshot in time.
• Neighborhood-level factors: neighborhood-level factors can explain at most 30% of the mobility gap, since the majority of the gap persists between black and white boys raised in the same neighborhood. In fact, the mobility gap tends to be greater in better neighborhoods. However, mobility gaps are smaller in areas with lower levels of poverty and higher levels of father presence.
• Individual-level factors: education explains much of the gap only among those with a college education or higher. But among blacks and whites with similarly low levels of education, there are still large gaps in mobility. Youth test scores explain most of the mobility gap across the entire distribution of test scores.

Implications

• Temporary income and wealth transfers, neighborhood/school integration, and affirmative action are unlikely to close much of the income gap in the long-run.
• Actions that successfully cause long-run increases in human capital, father presence, and self-control for blacks are likely to close most of the income gap within a few generations.

Racial disparities in socioeconomic outcomes

In this section, I document socioeconomic disparities between the major racial groups in the United States, particularly between black and white Americans. There are vast racial differences in income, wealth, and education that have been present for generations, many of which show no signs of closing. More importantly, there are also vast differences in mobility with respect to income and wealth, and education to a lesser degree, between blacks and whites.

Income

Black households have far lower average incomes than white households. US census data (2017) [archived] shows that the average black household income was more than one-third lower than the average white household income in 2017 ($40,258 vs $68,145, Table 1). This disparity has barely budged within over 50 years:

The same data shows that black households were over twice as likely to be poor as white households in the same year (21.2% vs 8.7%, Table 3). More recent data [archived] from the US census shows that, while the black poverty rate has decreased significantly since the 1960s, there are still large disparities in poverty rate between blacks and whites today:

Part of the explanation for racial disparities in household income is that black households are far less likely to be dual-income households than white households, independently of any income disparity between black and white individuals. However, data [archived] from the National Center for Education Statistics shows black workers have far lower annual earnings than white workers. The median annual earnings for black workers was $11,200 lower than the median earnings for white workers ($33,700 vs $44,900):

Controlling for educational attainment does not eliminate the disparity, as black workers have lower median earnings than white workers at each of the major levels of education:

Collins and Wanamaker (2017) [archived] have also provided data on racial differences in the distribution of income since 1900. Here’s the distribution of income score percentiles for black and white fathers and sons in 1900:

In 1973, there was almost no change in income distribution

In 1990, the income distribution for blacks had improved somewhat, but blacks were still disproportionately concentrated in the lower income brackets:

Wealth

In addition to income, there are also large black-white differences in wealth that have persisted for generations, as shown with data by the Pew Research Center [archived]. In fact, the ratio of median white wealth to median black wealth is greater today than it was in the 1980s:

Data from Brookings [archived] shows the wealth gap with more recent data:

Education

Data from Pew Research Center [archived] shows that the black-white disparity in high school completion has narrowed significantly in recent decades, although large gaps remain:

On the other hand, the college completion gap between blacks and whites has barely changed within 50 years:

Income mobility

The clearest illustration of racial differences in income mobility comes from Mazumder (2008) [archived]. Using the NLSY79, he examined income mobility by investigating the family income of subjects during adolescence (1978-1980) and during the ages of 32 to 44 (1997-2003). The rates of income mobility were reported as follows:

These graphs show stark disparities. Consider some of the following important findings from this data:

• Among children born into the bottom income quintile, blacks are nearly twice as likely to remain in the bottom income quintile as whites (44% vs 25%), and are less than half as likely to reach the top income quintile (4.1% vs 10.6%).
• Among children born into the top income quintile, blacks are nearly half as likely to remain in the top income quintile as whites (21% vs 39%), and are over twice as likely to end in the bottom income quintile (21.6% vs 10.4%). Surprisingly, blacks born in the top quintile are equally as likely to end in the bottom quintile as they are to remain in the top quintile (21.6% vs 21.3%).
• Blacks born in the middle quintile are nearly three times as likely to fall downward vs move upward (56% vs 22%), whereas the situation is reversed for whites: whites born in the middle quintile are more likely to move upward vs fall downward (43% vs 36%).
• Blacks born in the highest income quintile have similar outcomes as whites born in the second lowest income quintile. For example, blacks are more likely to end in the bottom quintile (21.6% vs 17.7%), about equally as likely to reach at least the middle income quintile (~60%), and about equally as likely to reach the top 2 quintiles (~40%).
• Blacks born in the second highest income quintile have similar outcomes as whites born in the bottom quintile. For example, blacks are more likely to end in the bottom income quintile (24.6% vs 24.9%), less likely to reach at least the middle income quintile (44.4% vs 50.7%), and less likely to reach the top quintile (7.7% vs 10.6%).

In addition to far lower rates of upward mobility, the data shows that blacks experience far higher rates of downward mobility than whites.

As can be inferred from the above data, black children are far less likely to exceed their parent’s income at every range of parental income:

Similar findings using more recent data was reported in Mazumder (2014) [archived] and Rodrigue and Reeves (2016) [archived].

Wealth mobility

Racial disparities in wealth mobility have been documented well by Pfeffer and Killewald (2019) [archived]. The researchers show extreme differences in wealth mobility by race, with black subjects being much more likely to be downwardly mobile and less likely to be upwardly mobile than white subjects. For example, the researchers note the following:

We display the estimated probability of attaining each quintile of the net worth distribution (dividing the wealth distribution into five equally sized groups) of the net worth distribution for black and white children who grow up in the same wealth quintile of the parental wealth distribution. For instance, among those growing up in the middle 20 percent of the parental wealth distribution, black children are much more likely to be downwardly mobile, with 39 percent of them falling to the bottom 20 percent of the wealth distribution compared to 16 percent of white children. Transition probabilities for each parental wealth quintile can be selected interactively and reaffirm the disadvantage of black children in attaining wealth irrespective of the wealth position of their parents.

The full data on the mobility rates is available in this animation [archived]. Using this data, I constructed the following graphs to show the percent of children reach each quintile by race and parental quintile:

Similar to income mobility, these graphs show stark racial disparities in wealth mobility. Consider the following surprising findings:

• There is very little difference in wealth between children born to parents in the lowest vs second lowest quintile. For example, black children born in either of these two quintiles have similar probabilities of falling to the bottom quintile (46% vs 39%), similar probabilities of reaching at least the middle quintile (27% vs 32%), and similar probabilities of reaching the top quintile (4% vs 5%).
• More shockingly, black children born to parents in the second highest wealth quintile have worse outcomes than white children born to parents in the lowest wealth quintile. For example, these black children are more likely to end in the bottom quintile (39% vs 32%), less likely to reach at least the middle quintile (32% vs 40%), and less likely to reach the top quintile (5% vs 7%).

These findings suggest that parental wealth does not benefit black children’s wealth as much as it benefits white children’s wealth. Indeed, this was reported directly by Pfeffer and Killewald (2017) [archived], who analyzed the same data. The researchers used data from the Panel Study of Income Dynamics (PSID) to study intergenerational patterns in wealth. Data was used from the earliest (1984-1989) and latest (2013-2015) wealth data collected in the study. The mean age of parents observed in 1984 is 43.4 years and 44.5 years for children in 2013. The mean intergenerational wealth correlation was 0.39, suggesting that “an advantage of 10 percentiles in the parent generation is associated with an advantage of 3.9 percentiles in the child generation” (page 15). However, they also find that “two-generational correlation in wealth positions is only half as strong for African Americans as for whites (0.18 versus 0.36)” (page 20). Table 5 shows the full data:

The authors then consider one possibility for why there is less intergenerational wealth persistence for the black subjects, namely that wealth persistence is greater at higher levels of wealth where black subjects are fairly absent. However, they reject this hypothesis since racial disparities in wealth mobility persist across all levels of wealth. They state the following (page 20):

One possibility is that the low two-generational wealth correlation observed for African Americans is due to their underrepresentation at the top of the wealth distribution, where persistence is greatest. Figure 1 shows race-specific mobility patterns across the wealth distribution. For every decile of the weighted parental wealth distribution, we plot the average wealth percentile of the offspring (ages 45–64) of those parents, separately by race…the figure also highlights that at each decile of parental wealth, the average wealth of African American adult children falls below that of their white peers. Thus, our results indicate that the higher intergenerational correlation in wealth for whites than African Americans is not exclusively due to African Americans’ severe underrepresentation at the very top of the distribution: race differences in mobility appear across the wealth distribution.

Educational mobility

There are also racial disparities in educational mobility, although these disparities are not as great as the income and wealth mobility gaps. Racial differences in educational mobility were studied by Ferrare (2016) [archived] using data from the General Social Survey’s (GSS) 1972-2014 cumulative data file. The GSS is a survey conducted by the National Opinion Research Center that uses a full probability sampling of households in the United States. The researchers investigate educational outcomes for subjects born between the 1910s to the 1980s. The following graphs show the probabilities of bachelor’s degree attainment at different levels of parental education by race, gender, and cohort (Figure 6):

As you can see, racial gaps appear to be greatest for most recent cohorts of children from parents with “high” education. For recent cohorts, this includes parents with at least 15-16 years of education. Both black men and black women appear much less likely to obtain a bachelor’s degree than their white counterparts, although the gaps border on statistical significance (see page 12) due to modest sample sizes and many slices of the data (cohort, gender, race, and parental education).

If we use reading and mathematics achievement as rough proxies for educational success, this may provide some insight into lower levels of educational mobility for black children. This can be examined using recent data [archived] on NAEP test scores (for the source: if using the archived webpage, click “Focus: racial/ethnic groups” on the left sidebar). In 2013, the following racial gaps in test scores by parental education were observed for 12th-grade students:

I’ve discussed this and other similar data in a separate post. This data is relevant for this post because it suggests that parental education does not benefit black children’s educational achievement as much as it does for white children. This can be noted by the fact that the gap in academic achievement increases among children with more educated parents. Consider the following findings for mathematics and reading scores, respectively:

• Mathematics: the black-white mathematics test score gap is 16 points, 24 points, and 32 points among children of parents who dropped out of high school, graduated from high school, and graduated from college, respectively.
• Reading: the black-white reading test score gap is 17 points, 25 points, and 31 points among children of parents who dropped out of high school, graduated from high school, and graduated from college, respectively.

Patterns of racial mobility gaps

In this section, I cite studies that analyze patterns of racial mobility gaps in more detail. The main points are as follows:

1. The black-white income mobility gap among males has remained fairly constant through most of the century and has reduced only slightly within the past few decades.
2. Controlling for parental household income, the black-white gap in individual income ranks is just one-third as large as the gap in household income ranks, likely because blacks are far less likely to marry even when conditioning on parental income.
3. When considering more racial groups, Asian Americans have the greatest mobility, followed by whites, then Hispanics, with blacks and American Indians having similarly low levels of mobility.
4. Many black-white mobility disparities are rather large for men but small (and sometimes non-existent or reversed) for women.

The black-white mobility gap throughout the 20th century

Before analyzing the mobility gap in detail, it is important to note the historical magnitude of the gap. This has been shown with data presented by Collins and Wanamaker (2017). These researchers document the gaps in mobility between black and white men from 1880 to 2000. They find that “conditional on father’s decile, there were stark racial differences in upward rank mobility in every cohort we can observe from the late nineteenth century through the end of the twentieth century” (page 16). More specifically, they note the following differences in mobility

In the historical samples, for those with the lowest earning fathers (1st decile), between 68 percent (1880-1900) and 85 (1910-1930) percent of white sons exceeded their father’s status compared to only 41 percent (1880-1900) or 59 percent (1910-1930) of black sons. The differences at the lowest two deciles in the early samples are both statistically significant and economically important; 97 percent of 1880 black fathers and 94 percent of 1910 black fathers were positioned in the two bottom deciles.

The basic pattern is similar for cohorts of men observed in the 1962 and 1973 OCG surveys. White sons exceeded black sons in upward rank mobility by about 20 to 30 percentage points at the bottom of the fathers’ rank distribution. From this perspective, there is no clear evidence that the first cohort of post-Civil Rights era black sons (outcomes measured in 1973) fared substantially better in terms of intergenerational mobility than those that preceded them.

For easier viewing, the authors also illustrates differences in mobility in Figure 3:

To quantify the magnitude of racial disparities in mobility, the researchers run regression analyses to estimate the association between race (being black) and income ranks among sons after controlling for father’s income rank. They find persistent gaps in mobility between black and white sons throughout, with the gap growing slightly throughout the middle of the 20th century followed by a slight reduction towards the end of the century. They described the findings as follows (page 18):

Panel A of Table 2 presents coefficients on an indicator variable for black sons in each cohort, conditioning only on the fathers’ rank. For sons observed in 1900 and 1930, black sons ranked between 22 and 23 percentiles lower than white sons in the “full sample” regressions; the gaps are even larger in the “30+ year old” samples. In the later cohorts, the racial mobility gaps were 27 percentiles in 1962, 25 in 1973, and 19 in 1990.

Table 2 shows the son’s income rank regressed on an indicator for black and father’s income rank:

In the conclusion of the paper, the researchers note that one of the most surprising findings is the “the mobility gap’s consistency over time, including through the Civil Rights Era, with its major changes in discriminatory policies and institutions” (page 24).

Offspring rank vs parent rank

Instead of showing the outcome income quintiles of children born into each income quintile, we can illustrate racial differences in mobility by showing mean rank outcomes of children born into each parental income rank. This exact analysis was performed by Chetty et al. (2020) [archived]. They note the following findings:

Figure II, Panel A plots the mean household income rank of children versus the household income rank of their parents, for black and white children. For whites, we estimate a slope (relative mobility) of βw = 0.32: a 10 percentile increase in parents’ rank is associated with a 3.2 percentile increase in children’s rank on average. The intercept for whites is αw = 36.8; that is, white children born to the lowest-income parents reach the 36.8th percentile on average.

The solid lines plot the mean child income rank at each parental income rank for whites and blacks. It may be useful to distinguish absolute mobility and relative mobility in this graph as the distinction appears several times in the paper.

• Absolute mobility refers to the mean rank of a child with parents at the lowest income rank. This determines whether the plotted line is shifted up or down in the graph. This is measured by the intercepts of the lines.
• Relative mobility refers to “the association between the mean percentile rank of children and their parents’ income ranks”. This determines the steepness of the plotted line in the graph. This is measured by the slope of the lines.

As you can see from the graph, blacks have far lower levels of absolute mobility (the black and white intercepts are 25.4 vs 36.8, respectively), but only slightly lower rates of relative mobility (the black and white slopes are 0.28 vs 0.32, respectively). The authors note this finding as well:

Blacks have relative mobility comparable to whites (βb = 0.28), but have uniformly lower rates of absolute mobility across the entire parental income distribution. For example, black children with parents at the 25th percentile reach an income rank of 32.6 on average, 12.6 percentiles below white children born to parents with comparable incomes. Racial disparities persist even at the highest income levels: among children whose parents are in the top 1% (who have incomes of $1.1 million on average), the black-white gap remains at 12.4 percentiles. Hence, high levels of parental income provide no insulation against racial disparities.

The researchers also illustrated the mobilities of other racial groups as well. They found that Asians have by far the highest rates of economic mobility, American Indians have mobility rates similar to blacks, and Hispanics have similar rates of absolute mobility as whites but lower relative mobility rates (Figure III-A).

Thus, even though Hispanics have income distributions “closer to those of blacks and Americans Indians than whites”, their high levels of mobility implies that “Hispanics are on an upward trajectory across generations and may close most of the gap between their incomes and those of whites”. Note that this implies that the descendants of current Hispanics are likely to close much of their gap in income ranks with whites. This does not imply that most of the overall income gap between Hispanics and whites will close in the future, since we will likely continue to receive more immigration of low-income Hispanics, who will likely have lower incomes than the descendants of current Hispanics.

As the graph above shows, comparisons between blacks and Asians reveals far larger gaps in mobility than comparisons between blacks and whites. The graph shows that Asians born in the lowest parental income rank attain higher mean income ranks than blacks born in the vast majority of parental income ranks. In fact, Asians born in the lowest income quintile are more likely to reach the highest income quintile than blacks born in the highest income quintile (Table 1), regardless of whether we use the children’s individual income (26.8% vs 26.2%) or household income (25.5% vs 18%) in adulthood:

Note, however, that almost all of the advantage for Asians is restricted to those with foreign-born parents. As the researchers note, “the exceptional outcomes of Asian children are unique to the children of first-generation immigrants rather than a persistent feature of Asians who are U.S. natives.” The authors note two possible explanations for why children of first-generation parents might have higher rates of mobility than the children of native parents. One possible explanation is that first-generation immigrants may have “low levels of earnings when they come to the United States despite having high levels of latent skills that they transmit to their children”. Another possibility is that “immigrants choose to live in areas within the United States that foster greater upward mobility for their children.”

When we graph the income mobility of all racial groups only for children with mothers born in the U.S., we find that Asians have slightly higher absolute mobility and slightly lower relative mobility than whites (Figure III-B):

Household income vs individual income

One explanation of the black-white intergenerational gap is that blacks are far less likely to marry than whites. This is because married households typically have much higher incomes than unmarried households. This is a plausible hypothesis because the authors show that blacks have far lower marriage rates than whites even when controlling for parental income (Figure IV-A).

We first document the large intergenerational gaps in marriage rates between black and white children in our sample. Figure IV, Panel A plots marriage rates for black and white children in 2015 (between ages of 32 and 37) by parental income percentile. Black children have substantially lower marriage rates across the parental income distribution, with a gap of 32 percentage points for children with parents at the 25th percentile and 34 percentage points at the 75th percentile. White children at the bottom of the income distribution are as likely to be married as black children at the 97th percentile of the parental income distribution.

Indeed, when one looks at child individual income rather than child household income, the black-white gap reduces significantly.

To evaluate the impacts of these differences in marriage rates, we focus on children’s individual incomes (excluding spousal income). Figure IV, Panel B plots children’s mean individual income ranks versus their parents’ household income ranks, by race. The gap in individual income ranks is approximately 5 percentiles across the parental income distribution, substantially smaller than the approximately 13 percentile gap in household income in Figure II, Panel A.

Note that the gaps in individual income ranks are much smaller than the gaps in household income ranks. Consider the following:

• Among children born in the 25th percentile of parental household income, the mean household income rank of black children was 12.6 percentiles lower than the same rank for white children (see Figure II from above). However, the mean individual rank for black children was 4.2 percentiles lower than the same rank for white children (see Figure IVb).
• Among children born in the 75th percentile of parental household income, the mean household income rank of black children was 15.6 percentiles lower than the same rank for white children (see Figure II from above). However, the mean individual rank for black children was 5.6 percentiles lower than the same rank for white children, (see Figure IVb).

So, among children from households with similar parental incomes, the gap in income ranks between black and white children is only about one-third as large when income is measured by individual rather than household income.

The income mobility gap by sex

This smaller gap in individual mobility hides stark gender differences in the racial gap in individual income. When analyzing the individual mobility gap for male and female children, the authors find that “the black-white intergenerational gap in individual incomes is driven almost entirely by men”

Figure V replicates Figure IV, Panel B separately for male and female children. This figure reveals that the black-white intergenerational gap in individual incomes is driven almost entirely by men. We find gaps for men of about 11 percentiles across the parental income distribution. In contrast, black women have 1 percentile higher individual income ranks than white women conditional on parental income.

Perhaps surprising to many, black females have slightly higher individual incomes than white females when controlling for parental income. Consider the following findings:

• At the 25th percentile of parental household income rank, the mean income rank for black males is 9.7 percentiles lower than that of white males, whereas the mean rank for black females is 1.4 percentiles greater than that of white females.
• At the 75th percentile of parental household income rank, the mean income rank for black males is 12 percentiles lower than that of white males, whereas the mean rank for black females is 1 percentile greater than that of white females.

Online Appendix Table V from the supplementary material [archived] repeats the analysis in Table 1 (see above) disaggregated by sex. Again, there are large gender differences in the racial mobility gaps:

Consider the following findings:

• Among black children born in the bottom quintile, males are nearly twice as likely as females to have individual incomes in the bottom quintile in adulthood (37.5% vs 20.5%). Similar patterns are observed for household incomes in adulthood.
• Among black children born in the top quintile, males are considerably more likely to have individual incomes in the bottom quintile in adulthood (16.4% vs 11%) and are nearly twice as likely to have household incomes in the bottom quintile in adulthood (21.5% vs 11.8%). In fact, black males born in the top quintile (21.5%) are more likely than a randomly selected child in the sample (20%) to have a household income in the bottom quintile in adulthood.
• Consistent with earlier data, income disparities are greatest when comparing blacks and Asians. For example, black males born in the top quintile are about equally as likely as Asian males born in the bottom quintile to end in the bottom quintile of individual income in adulthood (16.4% vs 17%).
• Blacks are the only racial group where the median income of females is greater than the median income of males.

More sex differences in black-white gaps

The above data shows large gender differences in racial gaps in income among children from households with comparable parental incomes. There are also similar gender differences in racial gaps in wage rates, hours worked, and employment rates. That is, among blacks and whites from households with similar parental incomes, black females perform about as well as (or better than) white females for these outcomes whereas, black males perform substantially worse than white males.

Figure VI plots mean wage ranks, hours, and employment rates by parental income percentile for women and men. Conditional on parental income, black and white women have very similar wage rates, hours of work, and employment rates…In contrast, there are very large gaps in both wage rates and hours of work for men. Conditional on parental income, black men have wages that are about 7 percentiles lower than white males, and work roughly nine fewer hours per week on average. The gaps in employment rates for men are particularly stark, especially for children growing up in low-income families. Black men with parents at the 25th percentile are 18.9 percentage points less likely to work in a given year than white men, whereas black men with parents at the 75th percentile are 11.4 percentage points less likely to work than white men. The employment rates of black men with parents at the 75th percentile are comparable to those of white men with parents at the 9th percentile.

These disparities also show large racial differences in gender dynamics. Among the white subjects, men have higher wages, work more hours, and have higher employment rates than women on average. These patterns are reversed for the black subjects: among the black subjects, men have similar wages as women, but lower hours worked and substantially lower employment rates than women.

Gender differences in racial gaps are also observed for other outcomes unrelated to the labor market, including educational attainment and incarceration. The authors find that, controlling for parental income, black females have comparable outcomes to white females regarding high school completion, college attendance, and incarceration. Black males, on the other hand, have substantially worse outcomes than their white counterparts for each of these outcomes.

We continue to find smaller intergenerational gaps for women and large intergenerational gaps for men for outcomes that are observed for everyone, such as educational attainment. Among children with parents at the 25th percentile, the black-white gap in high school completion rates is 3.5 percentage points for women versus 8.3 percentage points for men (Figure VII, Panels A and B). The corresponding gaps in college attendance rates are 2.8 percentage points for women and 6.5 percentage points for men (Figure VII, Panels C and D).17 It is particularly noteworthy that high school completion and college attendance rates are uniformly higher for black women than for white men across the parental income distribution.

As can be seen easily from the graphs, and as the authors have noted, the “gender difference in racial disparities is perhaps most stark in incarceration”. Panel E shows that “21% of black males born to parents in the lowest-income (bottom 1%) families were incarcerated” whereas only “6.4% of white males born to parents with comparable income were incarcerated”. Thus, black males born in the bottom 1% were over 3 times as likely as white males born in similar incomes to be incarcerated. Among children born in the top 1%, the authors again note large disparities: “only 0.2% of white males were incarcerated, whereas 2.2% of black males were incarcerated—the same rate as for white boys who grew up in families at the 34th percentile of the parental income distribution”.

I should note that gender differences in racial economic gaps have been present for decades. For example, Carlson and Swartz (1988) [archived] found that, among adults who reported positive earnings in in urban areas of 20 sample states in 1979, the annual earnings for black males was just 67% of the earnings of their white counterparts (Table 3), whereas the annual earnings for black females was about 98% of the earnings of their white counterparts (Table 4). The authors note that, even though black women had less education than white women, this disadvantage was “largely offset for black women by their greater hours worked” (page 540). These findings led Cunningham and Zalokar (1992) to conclude that the black-white pay gap among women had “almost disappeared” (page 546):

By 1980, in contrast to 1940 and 1960, the black-white pay gap [among women] had almost disappeared, and the remaining gap could be entirely explained by differences in measured characteristics. Racial differences in years of schooling completed were responsible for three-quarters of the remaining wage gap. This finding is consistent with the result of previous research that analyzed black women’s annual earnings in 1970 and 1980 and found that most of the black-white female pay gap in 1980 was due to differences in education, not discrimination, and that labor market discrimination against black women had apparently declined during the 1970s (Carlson and Swartz 1988).

Returning to the Chetty et al. paper, one important finding to note is that, despite little to no disparity in individual income mobility disparity between black and white women, the raw gap in individual income between black and white women will not close unless the gap between black and white men is also closed. This is because the parental household incomes for black females will continue to be lower than the parental household income for white females, because black fathers tend to either earn less money and/or not live in the household. Thus, black women will continue to earn less than white women, despite having similar rates of individual income mobility. The authors mention this same point:

We conclude based on the preceding analysis that the black-white intergenerational gap in individual income is substantial for men, but quite small for women. It is important to note, however, that this finding does not imply that the unconditional black-white gap in women’s individual incomes will vanish with time. This is because black women continue to have substantially lower levels of household income than white women, both because they are less likely to be married and because black men earn less than white men (Online Appendix Figure V). As a result, black girls grow up in lower-income households than white girls in each generation, leading to a persistent racial disparity in individual income for women even in the absence of an intergenerational gap in their individual incomes.

Thus, the authors note that “the key to closing income disparities for both black and white women is to close intergenerational gaps in income between black and white men”.

The importance of the black-white income mobility gap

In this section, I demonstrate why the income mobility gap between blacks and whites is important. The main points are as follows:

1. The income mobility gap, rather than income gaps of previous generations, is the primary proximate cause of the raw income gap.
2. The current income gap is in a steady state. If the current gap in mobility does not change, then the income gap is not expected to reduce any further.
3. If the racial gap in mobility was eliminated, the vast majority of the income gap would be eliminated within just two generations.

The income mobility gap is driving the raw income gap

The first reason that the income mobility gap is important is that the mobility gap is the main proximate cause of the raw income gap today. In other words, the income gap that we see today (and throughout the 20th century) exists primarily because of racial gaps in income mobility, rather than racial gaps in parental income.

One way to show this is by estimating counterfactual income distributions if blacks and whites had similar levels of mobility (while keeping actual gaps in parental income). Collins and Wanamaker (2017) performed this exact analysis for black and white fathers and sons throughout the 20th century. They found that, despite large initial gaps in income, “differences in transition patterns were the primary proximate cause of racial inequality in each generation” (page 23):

Given the racial differences in mobility documented above, it is natural to ask how far such differences go toward explaining differences in the distributions of sons’ labor market outcomes. For perspective, we estimate counterfactuals using re-weighting techniques developed in DiNardo, Fortin, and Lemieux (1996)….The result is a counterfactual for black sons that has fathers distributed as they are in the black sample but has the mobility patterns of whites. To be clear, this counterfactual is mechanical in the sense that it simply reshapes a given distribution of ranks. But it does provide quantitative perspective on how strongly the racial mobility gap has played into disparities in outcomes.

Figure 4 shows actual and counterfactual distributions of black sons’ rank. These kernel densities are smoothed distributions of the underlying data. In each panel, the solid lines, representing counterfactual outcomes for black sons, closely follow the long dashed lines representing actual outcomes of white sons. The result of equalizing mobility patterns is stark for all cohorts in the sample. Equality in transitions, conditional on fathers’ income ranks, would have greatly improved the outcomes of black men. The counterfactuals suggest that differences in transition patterns were the primary proximate cause of racial inequality in each generation.

These figures suggest that, since the 1930 cohort, if black sons had similar transition rates as white sons, black sons would have earned similar income distributions as white sons, despite large differences in parental income. In fact, these figures suggest that the income distributions of black males in the year 2000 was equivalent to the counterfactual income distributions that they would have reached in 1900 if they had similar income transition rates as white males (page 23):

It is not easy to imagine this counterfactual world of rapid black progress—obviously the course of American economic and social history would have been much different if, by 1900, black men had achieved the counterfactual distribution shown in Figure 4A. For reference, in census data, the median black male ranked at the 30th percentile of the national distribution of earned income in 2000. Thus, it took 100 additional years for the middle of the black male earnings distribution to reach the point that it might have reached in 1900 had black sons in our sample transitioned across income ranks like white sons with similarly ranked fathers.

The appendix [archived] attempts to precisely quantify the proportion of racial income disparities that are the result of racial differences in transition rates. To calculate this, they use two approaches to quantify the magnitude of the differences between two income distributions, one using a dissimilarity index and another using Hellinger distance estimates (read page 25 of the appendix for more detail on these approaches). Both measures indicate that the majority of racial income disparities for every cohort of males throughout the 20th century were the result of racial differences in transition rates. That is, using the aforementioned measures of differences between income distributions, they find that the counterfactual difference between white and black sons assuming similar rates of transition was less than half of the actual difference between white and black sons. See the following (page 25):

Our results indicate that mobility rates of black sons, conditional on their fathers’ income scores, comprise the majority of the differences in black and white sons’ relative positions in each cohort. These differences account for 66 to 80 percent of differences for the 1973 and 1990 cohorts and between 56 and 67 percent of the same in the historical samples through 1930.

The results for each cohort are presented here in the appendix:

As you can see, for any given cohort since 1930, the income disparities between black and white sons would have only been about 20-34% of the actual disparities if black and white sons had similar rates of income mobility. Note that this underestimates the gap reductions we would expect if black males had similar mobility rates as white males throughout the entirety of the 20th century. This is because, if black males had similar mobility rates as white males through the entirety of the 20th century, the gaps in parental income in the later cohorts would have been reduced as well. The gap reductions in the table are calculated using the actual parental income distributions for each cohort.

Based on this data, the authors “conclude that it was not only, or even primarily, initial income disparities per se that limited the pace of black workers’ economic progress in the historical samples. Rather, our results show a sharp disadvantage for black men relative to white men in the likelihood of escaping the bottom ranks of the income distribution throughout US history” (page 2) Instead, their findings suggest that “differences in transition patterns were the primary proximate cause of racial inequality in each generation” (page 23).

Current income disparities are in a steady state

The second reason that the income mobility gap is important is that the current gap in mobility suggests that current income disparities are in a steady state. In other words, if the current gap in income mobility does not reduce, then the income gap is not expected to reduce any further. In other words, in order to reduce the income gap, the mobility gap must first be reduced.

This finding has been shown by Chetty et al. (2020) [archived]. They used their data to predict steady-state mean income ranks for blacks and whites given current rates of mobility, i.e. the expected long-run income ranks of both blacks and whites assuming no changes in mobility. They estimate that the steady-state racial gap in mean income ranks is 19.2 percentiles, which is barely different from the actual gap of 20.9 percentiles. This suggests that “blacks and whites are in a steady state in which the black-white income gap is due almost entirely to differences in rates of intergenerational mobility rather than transitory or historical factors”:

Under the assumption that rates of mobility remain constant across generations, we can predict how the black-white disparity will evolve across generations using the model in Section II. Plugging our estimates of αw and βw into equation (3), the predicted steady-state mean rank for whites under the model in Section II is y¯SSw=54.4⁠, illustrated by the point where the intergenerational mobility line intersects the 45-degree line on Figure II, Panel A. The steady-state mean rank for blacks is y¯SSb=35.2⁠. Hence, the predicted (unconditional) black-white income gap in steady state given current levels of intergenerational mobility is Δy¯SS= 19.2 percentiles.

Figure II, Panel B plots the mean ranks of parents (circles) and children (diamonds) in our sample versus the predicted steady-state mean ranks, by race. Both blacks and whites’ mean incomes are close to their steady-state values, shown by the arrows intersecting the 45-degree line. The mean rank of black children in the 1978–1983 birth cohorts is 34.8, while the mean rank of white children is 55.7. Hence, the observed unconditional black-white gap in the current generation is 20.9 percentiles, very similar to the predicted steady-state gap of 19.2 percentiles. Interpreted using the model in Section II, this result implies that blacks and whites are in a steady state in which the black-white income gap is due almost entirely to differences in rates of intergenerational mobility rather than transitory or historical factors.

This has implications for policies that might be implemented to reduce income disparities between blacks and whites. For example, programs that temporarily transfer incomes from whites to blacks will not have much long-run impact on the income gap, because (assuming these programs do not address the mobility gap as well) blacks and whites will simply “revert” back to their steady states.

The large intergenerational gaps for blacks and American Indians relative to whites lead to disparities in earnings for these groups that persist across generations. If mobility rates do not change, our estimates imply a steady-state gap in family income ranks between whites and American Indians of 18 percentiles, and a white-black gap of 19 percentiles. These values are very similar to the empirically observed gaps for children in our sample, suggesting that blacks and American Indians are currently close to the steady-state income distributions that would prevail if differences in mobility rates remained constant across generations. This result shows that reducing racial disparities will require reducing intergenerational gaps—that is, disparities in children’s outcomes conditional on parental income—for blacks and American Indians. Transient programs that do not affect intergenerational mobility, such as temporary cash transfers, are insufficient to reduce disparities because income distributions will eventually revert back to their steady states.

Table I shows the steady state of black household income ranks in comparison to their actual income ranks. As you can see, the household income ranks for blacks and whites are predicted to converge to the 35th and 54th percentiles, respectively, in the steady state. For reference, the actual household income ranks for blacks and whites are currently at the 35th and 56th percentiles, respectively. In other words, the mean household income ranks for both blacks and whites are in their steady state and are therefore not expected to reduce any further (unless racial gaps in mobility are reduced).

The effects of eliminating the income mobility gap

The third reason that the income mobility gap is important is that, if the mobility gap were eliminated, the vast majority of the income gap would be eliminated within just two generations. I will show this using my own analysis in the next section. For the remainder of this section, I will reference Chetty et al. (2020) [archived] that shows this with their own analysis. Before doing so, recall from Figure II:

As you can see, the mean household income rank for black children is about 12 to 15 percentiles lower than the same rank for white children born to parents with similar incomes, depending on the specific level of parental income chosen. To estimate the expected income gap between blacks and whites in the counterfactual where they have similar rates of mobility, the authors estimate the predicted income gaps if black children’s mean household ranks were increased by 13 percentiles:

As noted in Section II, these steady-state predictions assume that mobility rates do not change across generations. Whether this assumption will hold going forward is unclear; the key point is that if the race-specific levels of absolute mobility αr do not change, there will be little progress in reducing black-white disparities in the United States. To reduce black-white disparities, we must reduce intergenerational gaps (Δα) either through changes in policy or other factors (e.g., via changes in ethnic capital as in Borjas 1992). Although the historical persistence of racial disparities suggests that reducing Δα will be challenging, one encouraging result is that interventions that reduce Δα could lead to rapid reductions in racial disparities across generations because blacks have fairly high rates of relative mobility (low βr). For example, under the assumptions of the model in Section II, if black children’s mean ranks were increased by 13 percentiles at all levels of parental income, the unconditional black-white income gap would fall to just 2.7 percentiles within two generations.

Recall that the current gap in mean household income ranks between blacks and whites is 20.9 percentiles (see Table I). The authors estimate that this gap would reduce to just 2.7 percentiles within two generations if blacks and whites had similar levels of mobility (i.e. if the mean household income ranks for black children were increased by 13 percentiles at all levels of parental income). This suggests that the gap in income ranks between blacks and whites would reduce to just 2.9/20.9 = 12.9% of its current size within just two generations if blacks and whites had similar levels of mobility. Thus, eliminating the mobility gap is not only necessary to close the income gap between blacks and whites, but it is probably also sufficient to do so within just a few generations.

Original analysis of black-white mobility gaps

In this section, I perform my own analysis of data available from other studies regarding the income and wealth mobility gap. The main points are as follows:

1. I reference an existing study that uses current rates of income mobility to estimate the steady-states of black and white income distributions. These income mobility rates suggest that contemporary income gaps can emerge within just 2-3 generations even if blacks and whites start with perfectly equal income distributions. Similarly, contemporary income gaps can vanish within just 2-3 generations if blacks had similar rates of income mobility as whites.
2. Using the same methodology to measure steady states, I also estimate the steady-states of black and white wealth distributions. Again, current race-specific wealth mobility rates suggest that contemporary wealth gaps can emerge within just 2-3 generations even if blacks and whites start with perfectly equal income distributions. Similarly, contemporary wealth gaps can vanish within just 2-3 generations if blacks had similar rates of wealth mobility as whites.

To calculate steady-state income/wealth distributions, I simply treat the distributions as Markov chains where each income/wealth quintile is a state and the quintile transition rates serve as transition probabilities. By simply pasting the black/white income/wealth transition rates into this calculator for solving Markov chains, we can estimate steady state income/wealth distributions for blacks and whites, given current mobility rates. To verify that this method is reliable, I calculate the steady state income distributions using transition rates presented by Mazumder (2012) [archived] and come to the same results as him.

Overall, the purpose of this section is to provide graphical illustrations of the fact that racial gaps in income/wealth mobility are proximately driven by gaps in mobility rather than the income/wealth distributions of previous generations.

Income

To start, I’ll estimate the steady-state of black and white income distribution using data presented by Mazumder (2012) [archived] using the calculator that I referenced above. Now, this author already calculates steady states for black/white income distributions by quintile, so my results here are not new. However, analyzing the data he has is useful because it confirms that the method I’m using to estimate steady states is compatible with current methods used in the literature.

To start, Table 7 displays the transition rates for black and white children:

This table is a bit difficult to read, so I converted the data to bar charts for easier viewing, consistent with the format of black/white income quintiles presented earlier:

Given the data provided above, there should be no surprises here. The clearest illustration of the differences here is the fact that blacks born to parents in the second highest income quintile have similar outcomes as whites born to parents in the lowest income quintile. Furthermore, blacks born to parents in the top quintile are equally as likely to reach the bottom quintile as whites born to parents in the second lowest quintile.

These numbers allowed the author to estimate the steady state distribution of income for blacks and whites. His results suggest that “Should these patterns of mobility persist, the implications for racial differences in the steady state distribution of income would be alarming. Instead of “regressing to the mean” as the standard IGE estimates would imply, these results would instead imply that blacks would make no further progress” (page 4). In particular the steady states for income distribution of blacks and whites were as follows (page 17):

The transition matrix of movements across quintiles of the income distribution over generations, for blacks and whites based on the SIPP-SSA are shown in Table 7…Assuming that these probabilities are a permanent feature of the US economy, they can be used to calculate an implied steady state distribution using standard matrix algebra methods for solving Markov chains. The results show, for example, that in the steady state, 39 percent of blacks would occupy the bottom quintile of the income distribution and only 8 percent would be in the top quintile. This suggests that rather than convergence, blacks would perpetually remain an underclass in American society if mobility patterns continue to evolve as they have for the cohorts studied in this paper.

The steady-state percent of blacks and whites in each other income quintile is mentioned in a footnote (page 18):

Further, 22 percent of blacks would be in the second quintile, 17 percent in the third quintile and 14 percent would be in the fourth quintile. The share of whites across the distribution, from the bottom to top quintiles is as follows: 17 percent, 20 percent, 20 percent, 21 percent and 22 percent.

Using the Markov chain calculator that I mentioned earlier (link here) and the transition rates mentioned in this study, I came to the exact same steady state estimates.

One interesting counterfactual to examine is the hypothetical where blacks and whites start with perfectly equal income distributions but maintain race-specific current rates of income mobility. This is actually fairly easy to calculate. If we know the transition rates for blacks and whites, then we can estimate exactly how many people are expected to transition to each quintile. For this counterfactual, I assume that, in the initial generation (G0), there are 20% of both blacks and whites in each income quintile. For each subsequent generation, I estimate the expected income distributions given the previous generation’s income distribution and the actual income mobility rates. The calculations and charts are all available in this Google spreadsheet. The findings were as follows:

As you can see, the perfectly equal income distribution (G0) is already fairly close to the steady state income distribution for whites. So their incomes do not change much through each generation. However, for blacks, there are rapid changes in income during the first two generations.

• After just one generation (G1), the black income distribution is already closer to the steady state distribution than to the original perfectly equal distribution. In G1, there are already nearly twice as many blacks (vs whites) in the bottom quintile and twice as many whites (vs blacks) in the top quintile.
• After just two generations (G2), the income distribution for blacks is already indistinguishable from their steady state distribution.

This shows just how rapidly differences in mobility can create large income disparities between two groups, even if the two groups have similar starting incomes.

Finally, another interesting counterfactual to examine is the hypothetical where blacks have the same mobility as whites today despite starting with current income distribution. This has already been illustrated in the Chetty at al. paper in the previous section. But it may be useful to illustrate the same point using other data and other kinds of illustrations (i.e. by showing how quintile distributions change rather than mean ranks change). For this counterfactual, I will set the income distribution for the starting generation (G0) of blacks equal to the current income steady state as reported by Mazumder (2012) [archived]. For each subsequent generation, I estimate the expected income distributions given the previous generation’s income and the white rates of income mobility. Again, the calculations and charts are all available in this Google spreadsheet. The findings were as follows:

As you can see, within just one generation, the black income distribution would be nearly perfectly equal if blacks had similar rates of mobility as whites, despite the fact that the starting income distribution is highly concentrated in the lower quintiles. Again, this shows that the proximate cause of income distributions are primarily determined by income mobility rather than the income distributions of past generations.

Wealth

In this section, I hope to perform the exact same analysis as the previous section, except for wealth instead of income. I already posted wealth data by Pfeffer and Killewald (2019) [archived] earlier in the post. For reference, I’ll repost the charts here:

Using the same methodology and Markov chain calculator mentioned in the previous subsection, I estimate steady states for wealth quintile distributions by race. I find that, in the steady state of wealth distribution, the share of whites across each quintile from bottom to top would be as follows: 16%, 19%, 21%, 22%, and 21%. The same figure for blacks is: 41%, 27%, 16%, 10%, and 6%. Unsurprisingly, in the steady state of wealth distribution, blacks are expected to possess far less wealth than whites.

As I did with income, I will now consider the counterfactual where blacks and whites start with perfectly equal wealth distributions but maintain current race-specific rates of wealth mobility. For this counterfactual, I assume that, in the initial generation (G0), there are 20% of both blacks and whites in each wealth quintile. For each subsequent generation, I estimate the expected wealth distributions given the previous generation’s wealth distribution and the actual wealth mobility rates. The calculations and charts are all available in this Google spreadsheet. The findings were as follows:

Again, the findings here are very similar to the findings for income. The perfectly equal wealth distribution (G0) is already fairly close to the steady state wealth distribution for whites, so their wealth does not change much through each generation. However, for blacks, we notice quite rapid changes in wealth during the first two generations.

• After just one generation (G1), the black wealth distribution is already closer to the steady state distribution than to the original perfectly equal distribution. In G1, there are already over twice as many blacks (vs whites) in the bottom quintile and over twice as many whites (vs blacks) in the top quintile.
• After just two generations (G2), the wealth distribution for blacks is already indistinguishable from their steady state distribution.

Again, this shows just how rapidly differences in mobility can create wealth disparities between two groups, even if the two groups have similar starting levels of wealth.

Finally, as I did with income, I also would like to examine the counterfactual where blacks have the same mobility as whites today despite starting with current wealth distribution. As far as I know, this has not been estimated by any works in the literature. For this counterfactual, I will set the wealth distribution for the starting generation (G0) of blacks equal to the current wealth steady states that were reported earlier in this subsection. For each subsequent generation, I estimate the expected wealth distributions given the previous generation’s wealth and the white rates of wealth mobility. Again, the calculations and charts are all available in this Google spreadsheet. The findings were as follows:

Just like with income, within just one generation, the black wealth distribution would be fairly evenly distributed if blacks had similar rates of mobility as whites, despite the fact that the starting wealth distribution is highly concentrated in the lower quintiles.

Explaining the black-white income mobility gap

In this section, I cite data from studies attempting to explain the income mobility gap between blacks and whites. These studies consider family-level factors (parental wealth, education, marital status), neighborhood-level factors, and individual-level factors (test scores, education) as possible explanations of the income mobility gap. The main results are as follows:

• Family-level factors: parental education and parental wealth explain very little (<20%) of the mobility gap. Family structure (i.e. marital status, single mother household) may explain a substantial portion of the upward mobility gap, although only if using measures of family structure throughout one’s entire childhood rather than a particular snapshot in time.
• Neighborhood-level factors: neighborhood-level factors can explain at most 30% of the mobility gap, since the majority of the gap persists between black and white boys raised in the same neighborhood. In fact, the mobility gap tends to be greater in better neighborhoods. However, mobility gaps are smaller in areas with lower levels of poverty and higher levels of father presence.
• Individual-level factors: education explains much of the gap only among those with a college education or higher. But among blacks and whites with similarly low levels of education, there are still large gaps in mobility. Youth test scores explain most of the mobility gap across the entire distribution of test scores.

Family-level factors

Chetty et al. (2020) [archived] considered a number of family-level factors as candidate explanations of the black-white mobility gap. They consider black-white gaps in income mobility after controlling for factors such as parental marital status, parental education, and parental wealth. They find that controlling for all of these variables together explain little of the income mobility gap. They note the following:

In Figure VIII, we show how Δp¯|X changes as we control for various factors X. Panel A considers the black-white gap for children growing up in low-income (⁠p¯=25⁠) families, while Panel B considers the gap for those growing up in high-income (⁠p¯=75⁠) families. As a reference, the first two bars in the panels report the unconditional difference in white and black children’s mean individual income ranks, without controlling for parental income or any other covariate. This unconditional gap is 17.6 percentiles for males and 4.8 percentiles for females. The second set of bars report estimates of Δp¯ when no controls Xi are included. These estimates correspond to the difference between the black and white series in Figure IV, Panel B at the 25th and 75th percentiles (under a linear approximation for both series).

These figures show how the black-white gap in children’s individual income ranks changes as we control for family- and neighborhood-level factors. The bars on the left in each pair report the black-white gap in individual income ranks for boys, while the bars on the right report the same statistics for girls. The first set of bars shows the unconditional black-white gap in mean individual income ranks. The second set of bars reports Δp¯⁠, the intergenerational gap in mean income ranks at percentile p¯ of the parental income distribution, estimated by regressing children’s income ranks on their parents’ ranks, an indicator for being white, and the interaction of these variables. Panel A reports estimates for p¯=25⁠, and Panel B reports estimates for p¯=75⁠.

As you can see, controlling for parental marital status, parental education, and parental wealth only slightly reduces the income mobility gap. Consider the following reductions visible from this data:

• Among children born to parents in the 25th percentile, the gap in mean individual income rank between black and white children is 10 percentiles. After controlling for parental marital status, education, and wealth, the gap reduces to 8 percentiles. So only 20% (2/10) of the gap was eliminated by controlling for these family-level variables.
• Among children born to parents in the 75th percentile, the gap in mean individual income rank between black and white children is 11.7 percentiles. After controlling for parental marital status, education, and wealth, the gap reduces to 10.2. So only 12.8% (1.5/11.7) of the gap was eliminated by controlling for these family-level variables.

In summary, only about 10-20% of the income mobility gap is eliminated after controlling for common family-level factors, with the exact amount varying depending on the level of parental income we analyze. These findings lead the authors to conclude “the family-level factors most commonly discussed in prior work explain very little of the differences in intergenerational mobility between black and white men.”

Fox (2016) [archived] also examined the possible role that parental wealth plays in explaining the black-white income mobility gap. She analyzed data from the PSID, a nationally representative, longitudinal survey following individuals and their offspring from 1968 to now. Mobility for both blacks and whites is measured with two metrics: upward mobility and downward mobility. Upward mobility is measured as the proportion of the population who exceed their parents’ income rank by at least 20 percentile points. Consistent with prior studies, she finds large differences in upward income mobility between black and white children. She then checks if wealth has an impact on this mobility gap by controlling for parental wealth. Shockingly, she finds that parental wealth is associated with higher rates of upward income mobility for whites, but not blacks (page 715):

Controlling for parental wealth, I find that higher wealth is associated with an increased likelihood of upward mobility for white families, but not black families. Figure 2 shows the likelihood of upward mobility at various points in the parental net worth distribution for children growing up in the bottom income quintile by race. Among children growing up in the bottom income quintile, white children whose parents had higher net worth were increasingly more likely to achieve upward mobility than similar black children. As parental net worth increases, the likelihood of upward mobility for white children increases. However, the same relationship between parental wealth and the likelihood of upward mobility does not exist for black children. There does not appear to be a positive association between parental wealth and the likelihood of upward mobility for black children. As a result, the black–white mobility gap actually increases as wealth increases (see Appendix Figure 5). At low levels of wealth, the likelihood of upward mobility for both black and white children is essentially the same.

She also checks if wealth is related to racial differences in downward income mobility, i.e. the percentage of subjects who fall below their parents’ income rank by at least 20 percentile points. While she does find a racial gap in downward income mobility (35% of whites versus 45% of blacks experience downward mobility), the gap is not statistically significantly, but she notes that this is “likely due to the small sample of black families in the top half of the income distribution” (page 718). Regardless, wealth does not seem to have an association with rates of downward mobility for either black or white families (page 718):

I find no conclusive evidence that parental wealth has a protective association with the likelihood of downward mobility for either black or white families (see Figure 4). Both the probit and lowess models do not predict any differences in mobility probabilities across the wealth distribution. In regard to the mobility gap, both models find the gap to be constant (and statistically insignificant) across levels of wealth. It is more likely that we would see a relationship between wealth and downward mobility if we restricted our analysis to the top 20th percentile versus the top half of the parental income distribution, but the sample of black families gets very small at the top of the distribution, so I follow previous research and only examine downward mobility from the top half.

Further analysis of the relationship between parental wealth and the black-white gap in upward income mobility reveals that total wealth levels do not explain most of the gap (page 720):

Using a Blinder–Oaxaca decomposition to explore the extent to which differences in upward mobility by race are due to differences in total net worth versus differential returns to wealth by race, I find that despite enormous wealth disparities between black and white families in the United States, most of the difference in mobility is due to differential returns to wealth as opposed to differences in wealth levels (see Appendix Table A6).

Table A6 from the appendix presents the following data:

In other words, if whites had the same wealth distribution as blacks, that would explain about 33% of the wealth gap, because there is a positive association between wealth and upward mobility for whites. However, if blacks had the same wealth distribution as whites, this would not explain a statistically significant proportion of the gap (in fact, the gap would grow by 13.2%, though this is not statistically significant), because there is no association between wealth and upward mobility for blacks. I believe that the latter figure is the more relevant figure because it suggests that any efforts to increase black wealth to match white wealth would do very little to reduce the racial gaps in upward mobility (on the other hand, the data suggests that one could reduce the racial mobility gap by about 33% if white wealth was decreased to match black wealth, but I doubt most would endorse this prescription).

Finally, consider Mazumder (2012) [archived] which also examined a number of family-level factors to explain the mobility gap. In contrast to Fox (2016), this study does find that wealth is modestly associated with higher rates of upward mobility for both blacks and whites, although only at the lowest levels of wealth. Regardless, the magnitude of the association between wealth and mobility is not great.

Figure 12 shows how the upward transition probability out of the bottom quintile varies over distribution of net worth in the SIPP-SSA sample. It is notable that in contrast to some of the other covariates, the pattern for wealth appears to be more nonlinear. For whites upward mobility rises with wealth in the bottom half of the wealth distribution but is fairly flat in the top half of the distribution. For blacks, there is a more striking upward slope at the bottom end of the wealth distribution and a similar leveling off in the middle of the distribution. Although the point estimates suggest a decline in upward mobility for the wealthiest blacks, this is driven by a small number of observations and is accompanied by very large confidence bands. Conditional on wealth, the black-white gap is about 20 percentage points or about 20 percent lower than the unconditional estimates. The fact that wealth only appears to matter at the bottom of the wealth distribution is consistent with the idea that wealth reflects borrowing constraints and that such constraints may inhibit upward mobility. 

One reason for the different findings in this study vs the Fox (2016) study may be that the latter examined rates of exceeding parents’ income rank by 20 percentiles, whereas this study mere examined rates of escaping the bottom quintile. Regardless, controlling for wealth seems to reduce only about 20% of the black-white gap in probability of transitioning out of the bottom quintile. As you can see, there are large gaps in mobility across the entire wealth spectrum.

The study also examined the impact of family structure on racial differences in mobility. Two different datasets were used to test this impact: the National Longitudinal Survey of Youth (NLSY79) and the Survey of Income and Program Participation (SIPP). Data from the NLSY suggests that “that family structure does not play much of a role in accounting for the black-white gap in upward mobility” (page 22). However, this may be because the NLSY79 dataset only includes data on whether the subject lived only with his mother at age 14. Controlling for this simplest measure did not do much to increase the probability of transitioning out of the bottom quintile for blacks. However, the SIPP employs a more robust measure of family structure, which contains the “entire marital history of parents over the child’s lifetime” (page 22). With this measure, the authors find that childhood family structure seems to explain a large portion of the mobility gap between blacks and whites (page 22):

In Figure 10 I compare the upward mobility rates for those sons who according to the SIPP always lived with both parents to those sons who ever lived with just a single parent. For whites upward mobility out of the bottom quintile actually declines slightly from 0.75 (0.02) for those who ever lived with just one parent to 0.71 (0.02) for those who always lived two parents. For blacks, however, we see an increase in the transition probability from 0.47 (0.02) to 0.58 (0.02). The black-white gap declines from 0.28 (0.02) to 0.13 (0.06). This 15 percentage point improvement in upward mobility for blacks relative to whites is statistically significant at the 5 percent level.

Surprisingly, the SIPP data suggests that family structure does not appear to have much of an impact on rates of downward mobility between blacks and whites. The author summarizes his findings on the impact of family structure on black-white gaps in mobility as follows (page 27):

Many commentators have pointed to prevalence of black children raised by single mothers as a source of racial gaps in economic success. I find supportive evidence that blacks raised in two parent families throughout childhood experience significantly greater upward mobility. Interestingly, family structure appears not to matter for whites or for rates of downward mobility for either blacks or whites.

Neighborhood-level factors

All data on neighborhood-level factors come from Chetty et al. (2020) [archived]. The authors analyze the relationship between region and black-white income mobility gaps using the following three different measures of region, in increasing order of granularity: commuting zones (which are aggregations of counties commonly used to define local labor markets), Census tracts (containing 4,256 people on average), and Census blocks (containing 50 people on average). Firstly, regarding commuting zones (CZs), they find that, while rates of mobility vary substantially across CZs, black boys have lower levels of mobility than white boys in virtually all CZs:

[T]here are substantial differences in black and white boys’ outcomes within virtually all CZs, for both children with parents at the 25th and 75th percentiles. Indeed, we find that the distributions of outcomes for blacks and whites across CZs are almost nonoverlapping. At the 90th percentile of the (unweighted) CZ-level distribution, black boys have a mean income rank of 45.1, which falls at the 16th percentile of the corresponding distribution for white boys. Black boys do not have the same prospects for upward mobility as white boys in virtually any CZ.

Next, the authors considered how well Census tracts and blocks, which are more granular areas, explain black-white gaps in mobility. They find that the racial income mobility gap is reduced by at most only 30% when black and white children are raised within the same neighborhood:

One of the most well-known explanations for the black-white gap is residential segregation: blacks and whites may have different outcomes because they tend to live in different neighborhoods (e.g., Massey and Denton 1993). To test this hypothesis, we include Census tract fixed effects in equation (5), effectively comparing the outcomes of children raised in the same neighborhood. Figure VIII, Panel A shows that including tract fixed effects reduces the black-white individual income gap among boys with parents at the 25th percentile (p = 25) from 10.0 percentiles to 7.7 percentiles. Indeed, even when we compare children who grow up on the same Census blocks (which contain 50 people on average) by adding block fixed effects, the intergenerational gap for boys remains at 7 percentiles at p = 25 and 7.9 percentiles at p = 75. In short, the vast majority of the black-white gap persists even among boys growing up in families with comparable incomes in the same neighborhood; differences in neighborhood quality explain at most 30% of the black-white intergenerational gap.

The finding that black children have lower incomes than white children with the same parental income in the same neighborhood is ubiquitous. In fact, after correcting for sampling error, the authors predict that “white boys have higher incomes in adulthood than black boys in 98.7% of tracts.” These findings also suggest that greater school integration is also unlikely to close the income mobility gap for low-income children:

These results imply that reducing residential segregation alone may be insufficient to close the black-white gap, since substantial disparities persist within neighborhoods. Moreover, because low-income children who live on the same block are likely to attend the same schools, simply enabling black and white children to attend the same schools, without creating greater racial integration within schools or making other changes that have differential effects by race, is also likely to be insufficient to close the gap.

The authors also analyzed a number of commonly used measures of neighborhood quality, such as e.g., poverty rates, test scores, educational attainment, etc. They examined the relationship between these measures and income mobility at the Census tract level. Unsurprisingly, many of these measures of neighborhood quality were positively associated with upward mobility. For example, there was a positive correlation between upward mobility rate and the share of the population above the poverty line. However, these correlations were typically stronger for white boys rather than black boys. As a result, these findings showed that ““good” neighborhoods tend to have larger intergenerational gaps between blacks and whites.” For example, Figure Xb shows that “the mean intergenerational gap increases by 2.5 percentiles when moving from the highest poverty neighborhoods to the lowest poverty neighborhoods”:

Because of these findings, the authors then investigate neighborhoods where black boys have high rates of mobility and black-white mobility gaps are small. They find that tracts in which low-income black boys reach incomes above the 50th percentile “almost all have poverty rates below 10%”, which is approximately the population-weighted median poverty rate across tracts within the United States. Among tracts with poverty rates below 10%, the authors then investigate aspects of neighborhood quality that predict smaller mobility gaps between blacks and whites.

In Figure XI, we correlate various tract-level characteristics with the black-white gap given parents at p = 25 (⁠Δy¯bw25⁠) to identify the characteristics of areas with smaller intergenerational gaps. In addition to the more traditional proxies for neighborhood quality considered above, we expand the set of tract-level characteristics we consider to include a set of race-specific measures—such as poverty rates for black and white families—as well as other variables that have differential effects by race, such as measures of racial bias.

Consistent with the findings from above, most measures of neighborhood quality (share college graduate, mean household income, etc.) are associated with larger racial mobility gaps, since they have greater correlations with income ranks for white boys than black boys. However, a few variables are associated with smaller mobility gaps, namely father presence, share of the population married, and share of the black population above the poverty line.

In particular, black father presence seems to be the strongest predictor of smaller mobility gaps between black and white boys (father presence is defined “as an indicator for whether the child is claimed by a male on a tax form in the year he or she is matched to a parent”). More specifically, the authors report that “the black-white intergenerational gap is 6.1 percentiles in tracts with the highest levels of black father presence, compared with 9.3 percentiles in the tracts with the lowest levels of father presence.” They also found that tracts with greater father presence had lower incarceration rates and higher employment rates for black boys. Moreover, the researchers also reported a number of interesting findings about the association between black father presence and outcomes for black boys:

• Black boys benefit from the presence of black fathers in particular, rather than black males in general. They note that “what matters is the number of black men involved in raising children in a tract, not the number of black men overall.”
• Black boys benefit from the presence of black fathers even when conditioning on the child’s parents’ marital status, suggesting that “the association with father presence is driven by a characteristic of the neighborhood in which the child grows up and is not simply a direct effect of the marital status of one’s own parents.”

To estimate whether the association between neighborhood quality and higher mobility rates is causal, the authors analyze how these associations interact with the age at which children move to better neighborhoods. For more detail on the methodology, see section VII.C. In short, the authors estimate that “children who move at birth to an area where we observe 1 percentile higher incomes for children of their race would pick up about 50% of that effect themselves through a causal effect of place.” In other words, about half of the association between neighborhood quality and income mobility seems to be causal. The authors conclude that the features of neighborhoods that predict smaller mobility gaps – e.g., black father presence and low poverty – do have causal effects, but that very few black children live in such neighborhoods:

We conclude that neighborhoods with low poverty rates, high rates of father presence among blacks, and low levels of racial bias among whites have better outcomes for black boys and smaller racial gaps. Examples of such areas include Silver Spring in Maryland and parts of Queens in New York, where black boys growing up in low-income (25th percentile) families rise above the national median on average in adulthood. But very few black families live in such places. Less than 5% of black children currently grow up in a Census tract with a poverty rate below 10% and more than half of black fathers present. In contrast, 63% of white children live in areas with poverty rates below 10% and more than half of white fathers present.

Individual-level factors

I will rely on analysis by Mazumder (2012) [archived] to measure the impacts of individual-level factors on the black-white mobility gap. Specifically, I will rely on his estimates of the change in the gap after controlling for educational attainment and test scores.

The impact of education on the mobility gap was mixed. he basic finding is that the mobility gap is almost entirely eliminated between college-educated blacks and whites, but remains rather large (sometimes larger than the unconditional gap) between blacks and whites with similarly low levels of education (page 20):

With respect to the racial gap in upward mobility, controlling for education provides something of a mixed picture. The racial gap in upward mobility among those with less than a high school education is actually higher than the unconditional estimate. On the other hand the racial gap narrows sharply with additional years of post secondary education. Indeed among those with 16 years of schooling the racial gap in upward mobility gap is essentially closed. Nevertheless, the racial gap is still quite large among those with some post-secondary education but who have not completed college. For example, the black white gap among those with 14 years of schooling is still sizable at 16 percent. Given that only 17 percent of blacks in the NLSY attained more than 14 years of schooling, this suggests that marginal improvements in educational attainment may not do a great deal to improve the overall upward mobility prospects of blacks.

On the other hand, controlling for AFQT test scores consistently eliminates the majority of the gap across all AFQT-score levels (page 21):

The effects of including one’s AFQT score on rates of upward mobility are shown in Panel E of Figure 9. Here the results provide a relatively clean and compelling story. For both blacks and whites upward mobility rises with AFQT scores in a fairly similar fashion. There are especially sharp gains in upward mobility associated with increases in test scores at the low end of the AFQT distribution. Upward mobility continues to rise at a somewhat slower but still strong rate in the middle and upper half of the AFQT distribution. Remarkably, the lines for blacks and whites are relatively close throughout the AFQT distribution. For example, the black-white gap in moving out of the bottom quintile is only 5.2 percentage points for those with median AFQT scores compared to the unconditional gap of 27 percentage points. This suggests that cognitive skills measured at adolescence can “account” for much of the black-white difference in upward mobility.

As Table 2 shows, the unconditional black-white gap in probability of transitioning out of the bottom quintile is 27 percentage points, with 48% vs 75% of whites expected to escape the bottom quintile. This gap reduces to just 5.2 percentage points among those with median AFQT scores, about 80% (1 – 5.2/27) lower than its original value. Consequently, the study concludes that “It is apparent that the cumulative effects of a variety of influences that affect cognitive ability by adolescence play a critical role in accounting for racial differences in upward and downward mobility” (page 26).

Similar findings were reported with slightly earlier versions of the same data in a Pew report by Mazumder (2008) [archived]. Prior to controlling for test scores, about 75% of whites raised in the bottom income quintile eventually transition out of that quintile, whereas only 56% of blacks do the same (Figure 3). However, among those with a median AFQT score, there is almost no difference in the likelihood of transitioning out of the bottom quintile: 81% for whites and 78% for blacks with median AFQT scores achieve this feat (page 30). The data suggests that “test scores can explain virtually the entire black-white mobility gap” (page 30). Note that, in a world of “perfect income mobility” where parental income had no association with offspring income, we would expect that 80% of those born in the bottom quintile would transition out of that quintile as adults. Thus, these findings show that black children born in the bottom quintile with just median AFQT scores are about as likely to escape the bottom quintile as would be expected in a world of “perfect income mobility”.

In an earlier version of their paper, Collins and Wanamaker (2017) [archived] also reported data on the association between test scores and the income mobility gap. They used two different datasets to perform this analysis: the NLSY79 and data from a subsample of World War II enlistees in 1943. As was done in the Mazumder (2012) study cited above, Collins and Wanamaker control for test scores in the NLSY79 dataset by using AFQT scores. The authors perform a counterfactual exercise where they estimate the mean income rank of black sons if they all had AFQT equal to the median for whites (page 26). They find that this reduces about half of the mobility gap:

For the 1990 cohort, we follow Bhattacharya and Mazumder and use reported AFQT scores as a proxy for human capital…A simple counterfactual exercise shows that raising AFQT scores for black sons from their documented level to the 90th percentile of the AFQT distribution for blacks would raise the predicted average income rank of black sons by 10 percentile points: from the 28th to 38th percentile of the income distribution, roughly half of the mobility gap measured in that year. Although this counterfactual may seem dramatic, the 90th percentile of the black AFQT distribution corresponds to the sample median for whites

So AFQT scores explain about 50% of the income mobility gap here. Why is this estimate much lower than the 80% gap reduction that I reported for Mazumder (2012)? There are a number of possibilities:

1. Mazumder analyzed subject incomes during the years 1997-2005 whereas Collins and Wanamaker analyzed incomes in 1990, and this may have played a role.
2. Mazumder reported the extent to which AFQT scores explained the racial gap in the probability of transitioning out of the bottom quintile, whereas Collins and Wanamaker reported the extent to which AFQT scores explained the gap in mean income ranks.
3. Collins and Wanamaker focused only on the income mobility gap of males, which is far greater than any mobility gap of females.
4. Collins and Wanamaker measured the role of AFQT scores on the income mobility gap after already controlling for years of education. This may actually diminish some of the impact of test scores since blacks receive more years of education with similar test scores (as I’ve shown in a separate post). In essence, if one measures the effect of test scores on income mobility while holding fixed years of education, this may give a downwardly biased estimate of the effect because it excludes any of the effect that is mediated through education (see here why controlling for mediators can create this bias).

Regardless, a 50% reduction in the gap in mean income ranks for males is a large reduction. The authors also estimated the impact of test scores on income mobility using Army General Classification Test (AGCT) scores for a subsample of World War II enlistees in 1943. Again, the authors perform a counterfactual exercise where they estimate the mean income rank of black sons if they all had AGCT scores equal to the 90th percentile of the black distribution. The authors find that this counterfactual would eliminate the vast majority (about 75%) of the racial mobility gap for this cohort (page 26):

A similar exercise for the 1930 cohort of sons gives consistent results. In this case, we follow Carruthers and Wanamaker (2017) and impute Army General Classification Test (AGCT) scores based on occupation-region-race categories and a subsample of World War II enlistees in 1943…A counterfactual exercise where black sons all obtain AGCT scores at the 90th percentile of the black distribution would more than double their predicted average income rank: from the 15th to 31st percentile of the income distribution, a change equal to roughly three-quarters of the racial mobility gap for this cohort.

Unfortunately, the authors do not indicate whether the 90th percentile of the black AGCT score distribution is equal to the median of the white AGCT score distributions (as was the case with the 1990 cohort). So it’s not clear whether this counterfactual indicates that 75% of the income mobility gap would be eliminated if blacks all had median scores for whites. Regardless, this provides further evidence that the income mobility gap is sensitive to human capital as measured by tests such as the AFQT or AGCT. Indeed, this leads the authors to conclude that human capital differences, as reflected in test scores, are a major proximate cause of the income mobility gap. In particular, the authors conclude that “in a proximate sense, our results suggest that differences in human capital accumulation, conditional on parents’ economic status, have underpinned a substantial portion of the black-white mobility gap throughout the twentieth century” (page 28).

So it is clear that differences in test scores “explain” a substantial portion of the black-white gap in income mobility (“explain” here is meant in a statistical, rather than causal, sense). However, before concluding this subsection, I would like to consider an argument from Chetty et al. (2020) against the hypothesis that racial differences in cognitive ability (as measured by scores on tests such as the AFQT) are causally responsible for the income mobility gap. They argue as follows:

The last family-level explanation we evaluate is the hypothesis that there are genetic differences in cognitive ability by race. Because we do not have measures of innate ability in our data, we cannot use the same approach as above to evaluate this hypothesis. However, two pieces of evidence suggest that differences in ability are unlikely to explain the intergenerational gaps we document. First, the prior literature suggests no ex ante biological reason that racial differences in cognitive ability would vary by gender (Rushton and Jensen 2005). Hence, our finding that black-white intergenerational gaps vary so sharply by gender casts doubt on ability as an explanation for the gaps we observe.

Second, most prior arguments for the ability hypothesis rest on the large gaps observed between black and white children on standardized tests (e.g., Hernstein and Murray 1994). However, black-white test score gaps do not vary significantly by gender. Data from the National Assessment of Educational Progress show that the black-white gap in test scores at age 9 for low-income (free- or reduced-price lunch–eligible) children is 0.48 std. dev. for boys versus 0.44 std. dev. for girls (Online Appendix Figure VIII). The fact that these test score gaps are not aligned with the earnings gaps across gender casts further doubt on the view that differences in cognitive ability, as measured by test scores, explain black-white gaps in earnings outcomes.

Now, these authors are arguing against the hypothesis that genetic differences in cognitive ability are responsible for the mobility gap. However, their argument can also be levied against the more general hypothesis that differences in cognitive ability (regardless of the causes of these cognitive differences, whether genetic, environmental, epigenetic, etc.) are responsible for the mobility gap. I believe that the argument that they provide is not sufficient reason to cast doubt on this hypothesis for a number of reasons:

1. Firstly, it could be the case that racial differences in cognitive ability (regardless of what causes the differences) are detrimental to both black males and females, but the disadvantage suffered by black females may be offset by other factors that compensate for their cognitive deficit. This possibility was explored by Neal (2004) [archived] who illustrated a number of factors that might result in higher earnings for black women relative to white women. For example, a disproportionate share of non-working white women are married women raising children with high-paying husbands, whereas a disproportionate share of non-working black women are single-mothers receiving government assistance (page 2). Thus, white non-working women likely have higher earning potentials than black non-working women. This difference in potentials may attenuate (or even reverse) the racial income gaps among working women that would be expected given the racial cognitive disparities. Also, the fact that black women often have worse marriage market prospects may create more incentive for black women to seek higher earnings  (page 14), which could also offset the effects of any cognitive deficits.
2. Secondly, low levels of cognitive ability might be particularly damaging for black males, since low cognitive ability predicts criminality and other antisocial behaviors particularly for men. Indeed, the data from Chetty shows that black males, but not black females, are much more likely to drop out of high school and be incarcerated than their white counterparts with similar parental incomes. Cognitive deficits might harm the economic prospects for black males in particular because it sets them on a pipeline to early antisocial behavior. Indeed, elsewhere I have cited data showing that controlling for AFQT scores alone eliminates most of the disparity in incarceration between black and white males.

Implications

In this last section, I consider (what I believe to be) important implications from this data regarding policies to close the income gap between blacks and whites. There are two main points here:

1. First, I consider actions that are unlikely to resolve much of the income gap in the long-run. This includes temporary income and wealth transfers, greater efforts for neighborhood/school integration, and affirmative action.
2. Next, I consider actions that are likely to resolve much of the income gap. This includes any actions that cause an increase in long-run human capital, father presence, and self-control for blacks.

Actions unlikely to close the income gap

In this subsection, I consider actions that are unlikely to close much of the income mobility gap in the long-run, which (as shown above) is necessary to close the raw gap in the long-run. I’ll consider temporary income transfers, temporary wealth transfers, affirmative action, and neighborhood/school integration.

Temporary income transfers are obviously not going to close the income gap in the long-run because they do nothing to address the mobility gap. If the mobility gap is not addressed, then blacks will simply revert back to their steady-state income distributions after such programs are enacted. Indeed, Chetty et al. (2020) interpreted their findings as suggesting that “policies focused on improving the economic outcomes of a single generation—such as cash transfer programs or minimum wage increases—can narrow the gap at a given point in time but are less likely to have persistent effects unless they also affect intergenerational mobility.”

Temporary wealth transfers are also unlikely to resolve the income gap in the long-run to a substantial degree since they are unlikely to resolve the income mobility gap. There are two reasons for this:

• There is little to no association between wealth and income mobility for black people. For example, Fox (2016) reported that “There does not appear to be a positive association between parental wealth and the likelihood of upward mobility for black children” (page 715). Also, Chetty et al. (2020) shows that introducing controls for wealth does very little to reduce gaps in income mobility after already controlling for parental education and marital status.
• There are also large gaps in wealth mobility. Thus, assuming there is some noticeable benefit to wealth on black income mobility (which does not seem to be the case), these benefits will dissipate within just a few generations after the wealth transfer because blacks will revert to their steady state (assuming differences in wealth mobility are not addressed).

Neighborhood and school integration are also unlikely to resolve much of the income mobility gap in the long-run. There are a few reasons for this:

• Chetty et al. (2020) find that “differences in neighborhood quality explain at most 30% of the black-white intergenerational gap” by comparing income mobility of blacks and whites children raised in the same Census blocks (which contain an average of 50 people). As Chetty et al. note, “these results imply that reducing residential segregation alone may be insufficient to close the black-white gap, since substantial disparities persist within neighborhoods”.
• Much of that 30% reduction cannot be assumed to be causal. In fact, causal analysis by Chetty suggests that about half of the association between neighborhood quality and youth outcomes is causal, suggesting that the other half is not causal.
• Regarding school integration, none of the studies directly examined school integration as a method to reduce the income mobility gap. But I’m inferring that such initiatives are unlikely to resolve the income mobility gap because, as Chetty et al. have shown, (a) the vast majority of the mobility gap remains between blacks and whites on the same block, and (b) “low-income children who live on the same block are likely to attend the same schools”.

Finally, affirmative action is unlikely to resolve much of the income mobility gap in the long-run. By affirmative action, I am specifically referring to racial preferences at universities that result in higher probabilities of admissions for black applicants despite holding equal all other qualifications. There are a few reasons for why I believe affirmative action will not resolve much of the income mobility gap:

• Chetty et al. (2020) shows that controlling for parental education has almost no effect on the income mobility gap (Figure VIII).
• The previous point should be sufficient, but the data I presented at the beginning of this post demonstrated a modest racial gap in educational mobility. Thus, any intended benefits of educational attainment conferred by affirmative action are less likely to be transmitted to one’s offspring. Of course, a permanent affirmative action scheme could avoid this particular problem.
• In a previous post, I documented evidence showing that affirmative action is only practiced by the most selective universities. Indeed, affirmative action only makes sense for the most selective universities, since most universities accept most applicants, which means black applicants don’t need racial preferences to be accepted to most universities. So the beneficiaries of affirmative action are probably black students who would have received a college degree even without racial preferences (perhaps at a less selective university). Since these black students in particular would have had similar income mobility rates as whites regardless of affirmative action (since college-educated blacks already have similar mobility rates as college-educated whites; Mazumder 2012, Figure 9), affirmative action is unlikely to have much impact on the overall income mobility gap. Instead, it serves to improve the economic standing of a relatively privileged, small share of black college graduates.

Actions likely to close the income gap

In this subsection, I will mention some actions that are likely to resolve much of the income mobility gap if enacted, which (as shown earlier) is sufficient to close the raw gap within just a few generations. The main actions I have in mind are actions that increase black human capital and father presence in the long-run, since the data suggest that these increases would likely have substantial improvements on the income mobility gap. In addition, increasing marriage rates and self-control among blacks may also benefit mobility rates as well.

Firstly, increasing human capital for blacks to the level of whites (or ideally even higher), as measured by tests such as the AFQT, is likely to resolve most of the income mobility gap. There are a few reasons to believe this:

• Mazumder (2012) showed that controlling for AFQT scores eliminated most (about 80%) of the racial gap in the probability of transitioning out of the bottom quintile. Collins and Wanamaker (2017) showed that controlling for AFQT scores eliminates about 50% of the gap in income mobility among black boys in particular. However, the true reduction may in fact be greater than 50% because, as stated earlier, this figure only concerns boys (who have more stubborn income gaps) and this figure is the reduction after already controlling for years of education (which will  mask any benefit of human capital that is mediated through years of education).
• There is plenty of data not cited here showing that most of the economic disparities between blacks and whites are eliminated after controlling for test scores. There is also strong evidence that human capital, as measured by standardized test scores, has a causal influence on economic outcomes such as educational attainment, occupational performance, and income.
• Analyzing the gaps between blacks and whites after controlling for individual test scores actually understates the likely impact of increasing black human capital. This is because increasing black human capital is beneficial not only because individuals directly benefit from having higher human capital, but also because individuals indirectly benefit by being raised in neighborhoods with higher average levels of human capital. As Chetty et al. (2020) showed, much (about half) of the association between neighborhood quality and better outcomes is causal. Thus, if increasing black human capital improves black neighborhoods with respect to mean incomes, poverty rates, incarceration rates, etc. (which is highly likely given data shown here), then this would also likely benefit black children independently of their individual level of human capital. This dynamic is very similar to the finding by Chetty et al. that being raised in a neighborhood with more fathers is beneficial to a child, independently of whether that child is raised with their father.

Secondly, increasing father presence for blacks to the level of whites (or ideally even higher) is likely to resolve a substantial portion of the income mobility gap. There are a few reasons to believe this:

• Cohabiting/married households have higher incomes than single/unmarried households, partially simply because dual households are more likely to have two income earners. For example, in 2006, the poverty rate for non-married black families was 5 times greater than the rate for married black families (35% vs 7%, see chart 13 in a 2010 report [archived] by the Heritage Foundation). So if income mobility is measured in terms of household income, increasing father presence is likely to improve income mobility. Indeed, recall that Chetty et al. showed that, controlling for parental household income, the black-white gap in individual income ranks is only one-third as great as the gap in household income ranks. This difference is likely attributable to the fact that black households are much less likely to be dual income households, which will exaggerate household income gaps independently of individual income gaps.
• Black children who are raised in households where two parents are always present have greater income mobility than children raised in households where only one parent is ever present (Mazumder 2012, Figure 9).
• Increasing black father presence rates would likely decrease the poverty rates of black neighborhoods. This is beneficial not just because children benefit from living in low-poverty households, but also because children benefit from living in low-poverty neighborhoods (see Chetty et al.).
• Finally, black children benefit from living in neighborhoods with higher father presence (see Chetty et al.).

As an aside, in addition to increasing father presence, there may be some additional benefit associated with increasing rates of marriage (or long-run, stable cohabitation). That is, not only are dual-parent households preferable to single-mother households, married or cohabiting men might be preferable to single men (even if the single men don’t have any children). Some reason to believe this may be beneficial comes from Chetty et al, since they find that black-white intergenerational gaps are smaller in neighborhoods with higher rates of marriage (see Figure XI). Furthermore, married/cohabiting men have better outcomes on a number of measures compared to single men. For example, married men tend to have higher incomes (Peake and Vandenbroucke 2019 [archived]) and lower criminal offending rates (Gottlieb and Sugie 2019 [archived]) than single men. Now, it’s not clear whether this association is causal. For example, some evidence suggests that marriage does have causal benefits (Sampson et al. 2006) whereas other evidence suggests that married men begin exhibiting lower rates of offending in the years prior to getting married (Lyngstad and Skardhamar 2013). But even if marriage itself has no causal benefits, there may nevertheless be benefits to promoting higher rates of marriage within the black community. For example, perhaps men who plan to marry/cohabitate are more likely to seek higher earnings, desist from crime, etc. If so, the promotion of marriage may be beneficial because it encourages men to plan to marry in their future, which may encourage them to exhibit beneficial behaviors that are desirable to potential spouses. Now, much of this is speculative but it is something worth considering.

Finally, in addition to increasing human capital and father presence (and possibly marriage/cohabitation as well), increasing self-control among black children is also likely key to reducing the income mobility gap. Self-control was not measured in any of the studies in this post, but I’ve cited data elsewhere showing that black children exhibit far higher levels of misbehavior and poor self-regulation early in life (source) and that childhood self-regulation is a good predictor for many important life outcomes (source). Thus, addressing racial differences in self-control and self-regulation is likely an important step to addressing racial differences in income mobility.

Solutions?

I think the evidence is strong that addressing racial disparities in human capital, self-control, and father presence are the keys to addressing racial gaps in income mobility. Now, how exactly do we enact these actions? How do we increase test scores, self-control, and father presence for black people in the long-run? That question is beyond the scope of this post and there may be no clear answer given the available data. So this post doesn’t have any tangible solutions that can be easily implemented with policy today. Regardless, any initiatives to address these disparities would need to meet the following conditions:

• Such initiatives need to start very early in life, because cognitive and behavioral racial disparities emerge within the first few years of life (as I’ve shown in this post).
• Such initiatives need to increase human capital, father presence, and self-control in the long-run. If blacks simply revert back to their current levels on these factors, then we cannot expect any long-run convergence of income mobility or raw income.

Now, as has been shown earlier, addressing racial disparities in income mobility is not only necessary to resolve the raw income gap in the long-run; it is also sufficient to resolve virtually all of the raw income gap within just a few generations. Thus, if one has the goal of closing income disparities between blacks and whites, it will be necessary to address disparities in human capital, father presence, and self-control. Moreover, once one addresses these disparities, there’s a good chance that this will be sufficient to close the income gap within just a few generations.

Main sources

• Mazumder (2012) [archived] studied black-white disparities in income mobility using two datasets: the National Longitudinal Survey of Youth (NLSY79) and Survey of Income and Program Participation (SIPP). Both datasets provide not only data showing racial differences in income mobility, but also allow for testing if plausible variables (e.g., education, test scores, single-mother household, etc.) can explain the racial gap. The NLSY79 focuses on the incomes of subjects during the years 1997, 1999, 2001, 2003 and 2005 when sample members were between the ages of 33 and 48. The SIPP examines the earnings from 2003 through 2007 of sons born sometime between 1959 and 1982, meaning sons were between 21 and 48 years old when earnings were observed.
• Collins and Wanamaker (2017) [archived] documented the income and occupational mobility of black and white American men from 1880 through 2000. For historical datasets (covering cohorts from 1880 to 1930), the researchers built new intergenerational datasets by linking census records from 1880 to 1900 and again from 1910 to 1930. The Occupational Changes in a Generation data (OCG) was used for sons’ outcomes collected in 1962 and 1973. Finally, the National Longitudinal Survey of Youth 1979 data (NLSY79) was used for sons’ outcomes in 1990 and 2000. A very brief summary was published by Brookings here.
• Chetty et al. (2020) [archived] studied racial disparities in intergenerational mobility from 1989 to 2015. They used data from the Census 2000 and 2010 short forms and data from federal income tax returns to cover nearly the entire population of the United States. This extensive data allowed analyzing the association between mobility gaps and a robust set of variables, including parental education, neighborhood, wealth, etc. A shorter version of the study is available in the brief here.
• This Google Spreadsheet contains all the calculations and charts that I used for the original analysis section of this post.