Changing Opportunity: Sociological Mechanisms Underlying Growing Class Gaps and Shrinking Race Gaps in Economic Mobility

When Black children constitute a small share of a county's population, they are more likely to interact with white peers.

“When Black children constitute a small share of a county's population, they are more likely to interact with white peers”

• cites Blau, 1977 sourceNot checked — no full text of the source
• cites Currarini, Jackson and Pin, 2009Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Cheng and Xie, 2013 sourcePartly supported

  The paper’s opening discussion of group proportion says that a group’s larger relative share increases opportunities for others to befriend its members. This is consistent with more white friendships where white peers are more prevalent, but the paper does not establish the stated relationship for Black children’s share of a county population: its empirical analysis concerns friendships among students within schools, not county-level interaction. (LLM, conf=0.84)

  Experimental — accuracy not yet measured

White-Black earnings gaps for children from low-income families reduced by 30%.

“Earnings increased for Black children at all parental income levels, reducing white-Black earnings gaps for children from low-income families by 30%.”

Among children who move at a young age (e.g., before age 8) to a community where parental employment increased between the 1978-1992 birth cohorts, earnings rise systematically from early to late birth cohorts.

“Consider children who move at a young age (e.g., before age 8) to a community where parental employment increased between the 1978-1992 birth cohorts.Among these children, earnings rise systematically as we move from early to late birth cohorts.”

The white class gap in average household incomes grew from $17,720 in the 1978 birth cohort to $20,950 in the 1992 birth cohort.

“grew from $17,720 in the 1978 birth cohort to $20,950 in the 1992 birth cohort.”

Conditional on employment rates for low-income white parents, the employment rates of Black parents or high-income white parents are not strongly related to the outcomes of low-income white children.

“Conditional on employment rates for lowincome white parents, the employment rates of Black parents or high-income white parents are not strongly related to the outcomes of low-income white children.”

The white-Black race gap in average household incomes for low-income families fell by 28%, from $20,810 for children born in 1978 to $14,910 for children born in 1992.

“the gap in average household incomes between white and Black children raised in low-income families fell by 28%, from $20,810 for children born in 1978 to $14,910 for children born in 1992.”

The magnitudes of the changes varied across counties, leading to considerable heterogeneity in trends across areas.

“However, the magnitudes of the changes varied across counties, leading to considerable heterogeneity in trends across areas.”

The correlation between changes in children's outcomes and changes in parental employment rates across subgroups is 0.91.

“The correlation between changes in children's outcomes and changes in parental employment rates across subgroups is 0.91.”

Children's outcomes are much more strongly related to parental employment rates of peers in their own birth cohort than adjacent birth cohorts.

“We find that children's outcomes are much more strongly related to parental employment rates of peers in their own birth cohort than adjacent birth cohorts”

Incomes in adulthood increased across all parental income levels for Black children.

“In contrast, incomes in adulthood increased across all parental income levels for Black children.”

The intergenerational persistence of household income ranks for white children increased by 28% between the 1978 and 1992 birth cohorts.

“The intergenerational persistence of household income ranks for white children increased by 28%.”

Community-level changes in the parents' generation-measured using a variety of parental outcomes-are highly correlated with children's outcomes in adulthood.

“community-level changes in the parents' generation-which can be measured using a variety of parental outcomes-are highly correlated with their children's outcomes in adulthood.”

Intergenerational mobility changed much more modestly for Hispanic, Asian, and American Indian children during the period studied.

“Intergenerational mobility also changed much more modestly for Hispanic, Asian, and American Indian children during the period we study.”

Class gaps grew and race gaps shrank similarly for nonmonetary outcomes such as educational attainment, standardized test scores, and mortality rates.

“Class gaps grew and race gaps shrank similarly for nonmonetary outcomes such as educational attainment, standardized test scores, and mortality rates.”

Changes in children's outcomes-earnings, SAT/ACT scores, and educational attainment-are strongly positively correlated with changes in parental employment rates across cohorts in their community, even controlling for the employment status of a child's own parents.

“changes in children's outcomes-earnings, SAT/ACT scores, and educational attainmentare strongly positively correlated with changes in parental employment rates across cohorts in their community, even when controlling for the employment status of a child's own parents.”

For Black children growing up in low-income families, the employment rates of low-income Black parents are generally far more predictive of outcomes than the employment rates of low-income white parents.

“Similarly, for Black children growing up in lowincome families, the employment rates of low-income Black parents are generally far more predictive of outcomes than the employment rates of low-income white parents.”

The outcomes of white children growing up in low-income families are primarily driven by the employment rates of low-income white parents.

“We find that the outcomes of white children growing up in low-income families are primarily driven by the employment rates of low-income white parents.”

The relationship between changes in children's outcomes and changes in parental employment rates is virtually identical across race and class groups.

“The relationship between changes in children's outcomes and changes in parental employment rates is virtually identical across race and class groups.”

There is a modest upward trend in the rank-rank slope when pooling racial groups.

“We find a modest upward trend in the rank-rank slope when pooling racial groups”

The correlation of income ranks increased from 0.23 in the 1978 birth cohort to 0.29 in the 1992 birth cohort.

“correlation of income ranks from 0.23 in the 1978 birth cohort to 0.29 in the 1992 birth cohort.”

In counties where Black children constitute a small share of the population, the outcomes of Black children are related to the employment rates of low-income white parents.

“In such counties, the outcomes of Black children are related to the employment rates of low-income white parents.”

Children's outcomes are most strongly related to parental employment rates of peers they are more likely to interact with, such as those in their own birth cohort.

“Children's outcomes are most strongly related to the parental employment rates of peers they are more likely to interact with, such as those in their own birth cohort”

The incomes of Black children increased across the parental income distribution.

“the incomes of Black children increased across the parental income distribution”

The outcomes of Black children are also related to the employment rates of low-income white parents in counties with higher rates of interracial marriage, controlling for racial shares.

“The outcomes of Black children are also related to the employment rates of low-income white parents in counties with higher rates of interracial marriage, a proxy for cross-racial interaction, controlling for racial shares.”

The correlation in mean income ranks across counties between children born to low-income parents in the 1978 and 1992 cohorts is around 0.8 for both white and Black children.

“the correlation in mean income ranks across counties between children born to low-income parents in the 1978 and 1992 cohorts is around 0.8 for both white and Black children.”

Intergenerational mobility changed rapidly by race and class in recent decades in the U.S.

“We show that intergenerational mobility changed rapidly by race and class in recent decades in the U.S. and study the causal mechanisms underlying those changes.”

The relationship between children's and parents' income ranks in our data is very similar to that reported by Chetty et al. (2014a) when pooling racial groups and measuring children's income at the same ages.

“The relationship between children's and parents' income ranks in our data is very similar to that reported by Chetty et al. (2014a) when we pool racial groups and measure children's income at the same ages.”

• cites Chetty et al. (2014a)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

The outcomes of white children with low-income parents deteriorated much more sharply in areas where employment rates for low-income white parents fell more.

“the outcomes of white children with low-income parents deteriorated much more sharply in areas where employment rates for low-income white parents fell more.”

Among children who made the same moves at older ages (e.g., after age 13), there is little difference in earnings as we move from early to late birth cohorts.

“In contrast, among children who made the same moves at older ages (e.g., after age 13), there is little difference in earnings as we move from early to late birth cohorts.”

Similar patterns of growing white class gaps and shrinking white-Black race gaps occurred in educational attainment and SAT and ACT scores.

“We find similar patterns of growing white class gaps and shrinking white-Black race gaps in educational attainment and SAT and ACT scores”

There was relatively little change across cohorts in the likelihood that white or Black children from families in the bottom income quintile reached the top quintile of the income distribution.

“By contrast, there was relatively little change across cohorts in the likelihood that white or Black children from families in the bottom income quintile reached the top quintile of the income distribution.”

Earnings increased for Black children at all parental income levels.

“Earnings increased for Black children at all parental income levels, reducing white-Black earnings gaps for children from low-income families by 30%.”

Non-monetary outcomes such as employment, incarceration, and mortality rates exhibit similar trends.

“Non-monetary outcomes, such as employment, incarceration, and mortality rates also exhibit similar trends.”

Between the 1978 and 1992 birth cohorts, earnings increased for white children from high-income families relative to white children from low-income families, increasing earnings gaps by parental income (class) by 30%.

“Between the 1978 and 1992 birth cohorts, earnings increased for white children from high-income families relative to white children from low-income families, increasing earnings gaps by parental income (”

Similar relationships exist between changes in children's outcomes and changes in other community-level characteristics, such as parental marriage rates and mortality rates.

“We find similar relationships between changes in children's outcomes and changes in other community-level characteristics, such as parental marriage rates and mortality rates.”

Outcomes deteriorated the most for low-income white families in areas that were historically better for these families, such as the Great Plains and the coasts.

“Outcomes deteriorated the most for low-income white families in areas that were historically better for these families, such as the Great Plains and the coasts.”

Black children born to parents at the 25th percentile of the national income distribution reached, on average, the 33.5th percentile in the 1978 cohort and the 35.1st percentile in the 1992 cohort.

“Black children born to parents at the 25th percentile of the national income distribution reached, on average, the 33.5th percentile in the 1978 cohort and the 35.1st percentile in the 1992 cohort.”

The paper finds similar relationships between changes in children's outcomes and changes in mean community-level parental income ranks measured when the child is age 27 or when the child is age 22.

“We find similar relationships between changes in children's outcomes and changes in mean community-level parental income ranks measured when the child is age 27 or when the child is age 22 (Appendix Figures A.24c,d).”

There is a strong positive relationship between the difference in siblings' outcomes and changes in parental employment rates across cohorts in the county to which they move.

“There is a strong positive relationship between the difference in siblings' outcomes and changes in parental employment rates across cohorts in the county to which they move.”

Economic mobility remained low in Atlanta, where children born to low-income Black families experienced virtually no change in mobility between 1978 and 1992.

“However, economic mobility remained low in Atlanta, where children born to low-income Black families experienced virtually no change in mobility between 1978 and 1992 (Appendix Table A.16).”

Column 5 of Table II shows that the estimates of β μ remain similar when controlling for deciles of the change in parental income before versus after the child's move.

“Column 5 of Table II shows that the estimates of β µ remains similar when we control for deciles of the change in parental income before versus after the child's move.”

The differences in β w and β b by race far exceed the differences one would predict simply based on the average racial mix of counties in which white and Black people live.

“The differences in β w and β b by race far exceed the differences one would predict simply based on the average racial mix of counties in which white and Black people live.”

Changes in children's outcomes are strongly positively correlated with changes in parental marriage rates.

“Changes in children's outcomes are strongly positively correlated with changes in parental marriage rates (Figure Vd) and negatively correlated with changes in parental mortality rates (Appendix Figure A.24f).”

For every race and class group, changes in children's outcomes across cohorts are strongly positively correlated with changes in parental employment rates in their community.

“For every race and class group, changes in children's outcomes across cohorts are strongly positively correlated with changes in parental employment rates in their community.”

The relationship between children's outcomes and adults' employment rates during their childhood is very similar to the relationship between children's outcomes and parental employment rates when children are 27 years old.

“Consistent with this interpretation, the relationship between children's outcomes and adults' employment rates during their childhood shown in Appendix Figure A.17 is very similar to the relationship between children's outcomes and parental employment rates when children are 27 years old (Appendix Figure A.19).”

In the 1992 birth cohort, Black children are more likely to be working in early adulthood than their white counterparts up to the 40th percentile of the national parental income distribution.

“In the 1992 birth cohort, Black children are more likely to be working in early adulthood than their white counterparts up to the 40th percentile of the national parental income distribution (Appendix Figure A.6).”

The paper finds similar relationships between changes in children's outcomes and other measures of community-level change.

“We find similar relationships between changes in children's outcomes and other measures of community-level change.”

Changes in parental education, wealth, occupation, and marital status each explain very little of the changes in intergenerational mobility for white children in recent cohorts.

“Changes in parental education, wealth, occupation, and marital status each explain very little of the changes in intergenerational mobility for white children in recent cohorts.”

Outcomes improved for children who grew up in communities with increasing parental employment rates.

“In particular, outcomes improved for children who grew up in communities with increasing parental employment rates, with larger effects for children who moved to such communities at younger ages.”

The declines were more modest in areas that were historically worse for these families, such as Appalachia and the industrial Midwest.

“The declines were more modest in areas that were historically worse for these families, such as Appalachia and the industrial Midwest”

Appendix Figure A.17 shows a clear positive relationship: children's incomes increased more across cohorts in counties where employment rates among adults of the same race during their childhood increased more.

“Appendix Figure A.17 shows a clear positive relationship: children's incomes increased more across cohorts in counties where employment rates among adults of the same race during their childhood increased more.”

• AssumptionThe authors state this as an assumption: Any measurement error or bias in children's and parents' incomes does not change across cohorts in ways that differ by race or class.
• AssumptionThe authors state this as an assumption: The composition of the sample does not change across cohorts in ways that differ by race or class.

Estimates controlling for either childhood county or childhood Census tract fixed effects are essentially identical to our baseline estimates.

“The final two pairs of bars in Figure IV show that estimates controlling for either childhood county or childhood Census tract fixed effects are essentially identical to our baseline estimates.”

For children born in early cohorts (1978-1983), the relationship between age at move and outcomes is much flatter.

“For those born in early cohorts (1978)(1979)(1980)(1981)(1982)(1983), the relationship is much flatter, as expected because these cohorts do not experience the improvements in parental employment rates that occurred for subsequent cohorts.”

Economic outcomes deteriorated sharply for white children from low-income families relative to white children from high-income families in recent birth cohorts in the United States.

“This paper has shown that economic outcomes deteriorated sharply for white children from low-income families relative to white children from high-income families in recent birth cohorts in the United States.”

By the 1992 birth cohort, Black children growing up in families at the 25th percentile of the national income distribution were 3.3 percentage points more likely to be employed at age 27 than their white counterparts.

“By the 1992 birth cohort, this employment gap had reversed, with Black children growing up in families at the 25th percentile of the national income distribution 3.3 percentage points more likely to be employed than their white counterparts (Figure IIa).”

The correlation between the baseline measure of parental employment and the cross-cohort measure is stronger in all subgroups than the correlation with the cross-year measure.

“we find that the correlation between our baseline measure and the cross-cohort measure is stronger in all subgroups than the correlation with the cross-year measure (Appendix Figure A.26).”

In counties with low rates of interracial marriage, changes in Black children's outcomes are unrelated to changes in white parents' employment rates (β0w = -0.01).

“Holding fixed exposure to white people at the mean, changes in Black children's outcomes are strongly predicted by changes in white parents' employment rates in counties with high rates of interracial marriage ( β 1 w = 0.23), but are unrelated to white parents' employment rates in counties with low rates of interracial marriage ( β 0 w = -0.01).”

The share of childhood Facebook friendships decays rapidly with distance from one's own birth cohort.

“The share of friendships decays rapidly with distance from one's own birth cohort; intuitively, children are more likely to meet and befriend children in their own grade in school.”

Placebo estimates predicting children's household income ranks using parental income in early childhood (ages 0-4) and other pre-move parental characteristics are statistically indistinguishable from zero.

“In Columns 6 and 7, we predict children's household income ranks, ŷi , using parental income in early childhood (when the child was ages 0-4) (in Column 6) and additionally using other pre-move parental characteristics-education, occupation, wealth, and marital status (in Column 7).These placebo estimates are statistically indistinguishable from zero.”

Among children taking the SAT/ACT, the white class gap in mean SAT/ACT scores grew by 62% between the 1980 and 1997 birth cohorts.

“Among children taking the SAT/ACT, we find an especially large increase in the white class gap in mean SAT/ACT scores, which grew by 62% between the 1980 and 1997 birth cohorts.”

Changes for Hispanic, Asian, and AIAN children were generally much more modest than the changes for white and Black children.

“Changes for the Hispanic, Asian, and AIAN children were generally much more modest than the changes for white and Black children.”

The paper obtains a much flatter slope of 0.04 for children who move late in childhood (ages 13-17), indicating that the gains from moving to a community with increasing parental employment rates are much smaller for children who move at older ages.

“We obtain a much flatter slope of 0.04, indicating that the gains from moving to a community with increasing parental employment rates are much smaller for children who move at older ages.”

The white class gap increased by 3.37 percentiles between the 1978 and 1992 birth cohorts in the baseline specification with no controls.

“The first orange bar reports our baseline estimates with no controls, and shows that the white class gap increased by 3.37 percentiles between the 1978 and 1992 birth cohorts.”

• AssumptionThe authors state this as an assumption: Any measurement error or bias in children's and parents' incomes does not change across cohorts in ways that differ by race or class.
• AssumptionThe authors state this as an assumption: The composition of the sample does not change across cohorts in ways that differ by race or class.

For white children with low-income parents growing up in communities where parental employment remained more stable, mean household income ranks did not change meaningfully.

“In contrast, for those growing up in communities where parental employment remained more stable, mean household income ranks did not change meaningfully.”

Outcomes for Black children improved across the parental income distribution.

“Outcomes for Black children improved across the parental income distribution.”

These correlations increase to 0.87 for low-income white families and 0.78 for low-income Black families when comparing the 1978-1980 versus 1990-1992 birth cohorts to reduce measurement error.

“These correlations further increase to 0.87 for low-income white families and 0.78 for low-income Black families when we compare the 1978-1980 versus 1990-1992 birth cohorts to reduce measurement error.”

The paper's findings are not sensitive to the time point at which parental employment rates are measured.

“We show below that our findings are not sensitive to the time point at which parental employment rates are measured because most of the changes in parental employment rates are driven by differences across cohorts that are stable over time-and therefore likely reflect latent factors that were present during childhood-rather than year-specific shocks.”

Between the 1978 and 1992 birth cohorts, the gap in incomes between white children raised in low- versus high-income families grew (growing white class gaps).

“between the 1978 and 1992 birth cohorts, the gap in incomes between white children raised in low-versus high-income families grew-a pattern we term growing white class gaps”

• AssumptionThe authors state this as an assumption: Any measurement error or bias in children's and parents' incomes does not change across cohorts in ways that differ by race or class.
• AssumptionThe authors state this as an assumption: The composition of the sample does not change across cohorts in ways that differ by race or class.

Children are much more likely to interact with peers in their own birth cohort than in adjacent cohorts.

“Children are much more likely to interact with peers in their own birth cohort than in adjacent cohorts.”

The relationship between children's outcomes and parental employment rates is virtually identical across the three subgroups (white low-income, Black low-income, white high-income).

“Importantly, the relationship between children's outcomes and parental employment rates is virtually identical across the three subgroups.”

In the 1978 birth cohort, Black children growing up in families at the 25th percentile of the national income distribution were 2 percentage points less likely to be employed at age 27 compared to their white counterparts.

“in the 1978 birth cohort, Black children growing up in families at the 25th percentile of the national income distribution were 2 percentage points less likely to be employed at age 27 compared to their white counterparts.”

In counties with high rates of interracial marriage, changes in Black children's outcomes are strongly predicted by changes in white parents' employment rates (β1w = 0.23), holding fixed exposure to white people at the mean.

“Holding fixed exposure to white people at the mean, changes in Black children's outcomes are strongly predicted by changes in white parents' employment rates in counties with high rates of interracial marriage ( β 1 w = 0.23), but are unrelated to white parents' employment rates in counties with low rates of interracial marriage ( β 0 w = -0.01).”

The younger sibling has better outcomes on average than the older sibling when the family moves to a community that is improving across cohorts.

“The younger sibling has better outcomes on average than the older sibling when the family moves to a community that is improving across cohorts.”

• Also stated in: § E Sibling Comparisons: Estimator and Results · plain

  “The younger sibling has better outcomes on average than the older sibling when the family moves to a community that is improving across cohorts.”

For children born in later cohorts (1987-1992) who are exposed to increasing parental employment rates in the destination, there is a steep downward-sloping relationship: moving at an earlier age to an area with increasing parental employment rates is highly beneficial.

“For children born in later cohorts (1987)(1988)(1989)(1990)(1991)(1992) who are exposed to increasing parental employment rates in the destination, there is a steep downward-sloping relationship: moving at an earlier age to an area with increasing parental employment rates is highly beneficial.”

The magnitudes of these changes varied substantially across areas.

“However, the magnitudes of these changes varied substantially across areas.”

The white-Black race gap decreased by 4.16 percentiles between the 1978 and 1992 birth cohorts in the baseline specification with no controls.

“The first green bar reports our baseline estimate with no controls, and shows that the white-Black race gap decreased by 4.16 percentiles between the 1978 and 1992 birth cohorts.”

• AssumptionThe authors state this as an assumption: Any measurement error or bias in children's and parents' incomes does not change across cohorts in ways that differ by race or class.
• AssumptionThe authors state this as an assumption: The composition of the sample does not change across cohorts in ways that differ by race or class.

Changes in children's outcomes are negatively correlated with changes in parental mortality rates.

“Changes in children's outcomes are strongly positively correlated with changes in parental marriage rates (Figure Vd) and negatively correlated with changes in parental mortality rates (Appendix Figure A.24f).”

Charlotte (Mecklenburg County, NC) and Atlanta (Fulton County, GA) both had similarly low rates of economic mobility for children born in 1978, especially for children born to low-income Black families.

“Charlotte (Mecklenburg County, NC) and Atlanta (Fulton County, GA) both had similarly low rates of economic mobility for children born in 1978, especially for children born to low-income Black families.”

Across the fifty most populous counties, the correlation between upward mobility in the 1978 versus 1992 birth cohorts is 0.75 for white children and 0.62 for Black children.

“Across the fifty most populous counties, the correlation between upward mobility (the mean income rank in adulthood of children raised in low-income families) in the 1978 versus 1992 birth cohorts is 0.75 for white children and 0.62 for Black children (Appendix Figure A.16).”

From the 1978 to 1992 birth cohorts, children's incomes in adulthood fell for low-income white families in nearly every part of the country.

“From the 1978 to 1992 birth cohorts, children's incomes in adulthood fell for low-income white families in nearly every part of the country.”

The paper finds a similar relationship between changes in children's outcomes and changes in parental employment rates when measuring parental employment at different points in time.

“We find a similar relationship between changes in children's outcomes and changes in parental employment rates when measuring parental employment at different points in time.”

Over three-quarters of Black children are born to parents with below-median incomes.

“over three-quarters of Black children are born to parents with belowmedian incomes (Table I).”

• cites Table INot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

The decay of the coefficients on parental employment rates across adjacent cohorts closely mirrors the pattern of decay of social interaction across grades.

“Furthermore, the decay of the coefficients on parental employment rates across adjacent cohorts closely mirrors the pattern of decay of social interaction across grades.”

Children's outcomes are strongly related to parental employment rates of other children in their own birth cohort.

“Children's outcomes are strongly related to parental employment rates of other children in their own birth cohort.”

The effects of growing up in communities with increasing parental employment rates are larger for children who moved to such communities at younger ages.

“In particular, outcomes improved for children who grew up in communities with increasing parental employment rates, with larger effects for children who moved to such communities at younger ages.”

The relationship between parental employment rates and children's outcomes across cohorts is much flatter for late-childhood movers than for young movers.

“This relationship is much flatter across cohorts than the corresponding series for young movers, implying that increases in parental employment rates translate to much smaller changes in children's outcomes if they arrive in the destination county late in childhood.”

Children's outcomes are more strongly related to the parental employment rates of peers they are more likely to interact with.

“Children's outcomes are more strongly related to the parental employment rates of peers they are more likely to interact with, suggesting that social interaction mediates changes in economic mobility.”

Most of the changes in parental employment rates are driven by differences across cohorts that are stable over time rather than year-specific shocks.

“Our findings are not sensitive to the calendar year in which we measure parental employment because most of the variation in parental employment rates arises from differences across cohorts rather than across calendar years.”

Measuring parental employment in a fixed calendar year (e.g., 2012 or 2019) for all cohorts yields similar results.

“For example, measuring parental employment in a fixed calendar year (e.g., 2012 or 2019) for all cohorts-and hence at different child ages-yields similar results (Appendix Figures A.24a,b).”

The correlations with parental marriage and mortality rates hold both across subgroups and within subgroups across area.

“As with parental employment rates, these correlations hold both across subgroups and within subgroups across area.”

There is no relationship between changes in parental employment rates and predicted outcomes across cohorts and move ages.

“Figure VIId shows non-parametrically that there is no relationship between changes in parental employment rates and predicted outcomes across cohorts and move ages.”

Black children raised in low-income families and white children raised in high-income families fared worse if they grew up in communities with falling parental employment rates.

“Black children raised in low-income families and white children raised in high-income families also fared worse if they grew up in communities with falling parental employment rates.”

Between the 1978 and 1992 birth cohorts, the gap in incomes between white versus Black children raised in low-income families fell (shrinking white-Black race gaps).

“while the gap in incomes between white versus Black children raised in low-income families fell-a pattern we term shrinking white-Black race gaps.”

• AssumptionThe authors state this as an assumption: Any measurement error or bias in children's and parents' incomes does not change across cohorts in ways that differ by race or class.
• AssumptionThe authors state this as an assumption: The composition of the sample does not change across cohorts in ways that differ by race or class.

Between the 1978 and 1992 birth cohorts, the white class gap in four-year college graduation rates changed by 29%.

“White Class Gap (∆ = 29%) White-Black Gap at P=25 (∆ = -8%)”

For white children at P=75, the slope of the change in child household income percentile (1978-1992) on baseline child household income percentile (1978 cohort) is -0.37 (SE 0.02).

“Slope = -0.37 (0.02)”

In publicly available ACS data, the white-Black race gap in mean household income percentile at age 27 narrowed by 2.0 percentile points between the 1978 and 1992 birth cohorts.

“Gap = 17.8 Gap = 15.8Δ White-Black Gap = -2.0”

For white children at P=75, a 1 percentage point increase in parental marriage rate is associated with a 0.471 increase in child household income rank (SE 0.019) in the baseline specification.

“△ Parental Marriage Rate 0.471 0.423 0.427 0.420 (0.019) (0.020) (0.024) (0.025)”

By the 1992 birth cohort, the white-Black race gap in remaining in the bottom household income quintile shrank to 4.1 percentage points, a 72% reduction.

“By the 1992 birth cohort, the white-Black race gap shrank to 4.1 percentage points-a 72% reduction.”

The white-Black race gap in children's mean household income ranks in adulthood decreased by 27% between the 1978 and 1992 birth cohorts, from 14.9 to 10.9 percentiles.

“while the white-Black race gap decreased by 27%, from 14.9 to 10.9 percentiles.”

The weighted correlation between national-level changes in children's household income ranks and national-level changes in parental mortality rates is -0.93.

“Slope = -1.86(0.27) Correlation = -0.93”

For white children at P=25, a 1 percentage point increase in parental employment rate is associated with a 0.321 increase in child household income rank (SE 0.017) in the baseline specification.

“△ Parental Emp.Rate 0.321 0.287 0.126 0.096 (0.017) (0.017) (0.019) (0.019)”

Between the 1978 birth cohort and the last available birth cohort, the white-Black race gap in marriage rates at P=25 at age 32 changed by -19%.

“White Class Gap (Δ = 35%) White-Black Gap at P=25 (Δ = -19%)”

For white children at P=75, a 1 percentage point increase in parental employment rate is associated with a 0.322 increase in child household income rank (SE 0.024) in the baseline specification.

“△ Parental Emp.Rate 0.322 0.171 0.177 0.026 (0.024) (0.022) (0.029) (0.028)”

Between the 1978 birth cohort and the last available birth cohort, the white-Black race gap in marriage rates at P=25 at age 27 changed by -33%.

“White Class Gap (Δ = 16%) White-Black Gap at P=25 (Δ = -33%)”

In the 1978 birth cohort, Black children born to families in the bottom household income quintile were 14.7 percentage points more likely to remain in the bottom quintile than their white counterparts.

“In the 1978 birth cohort, Black children born to families in the bottom household income quintile were 14.7 percentage points more likely to remain in the bottom quintile than their white counterparts”

In ACS data, the white-Black gap for low-education male adults in employment rates changed by -0.3 percentage points (-2%).

“White High-Educ.vs. Low-Educ.Gap Δ = 1.9pp (17%) White-Black Gap for Low-Educ.Δ = -0.3pp(-2%)”

In the baseline sample from linked Census and tax records, the white-Black race gap in mean household income percentile at age 27 narrowed by 1.7 percentile points between the 1978 and 1992 birth cohorts.

“Gap = 19.4Gap = 17.6 Δ White-Black Gap = -1.7”

The white-Black race gap at P=25 in child individual income percentile decreased by 35% between the 1978 and 1992 birth cohorts.

“White-Black Gap at P=25 (∆ = -35%)”

In ACS data, the white-Black gap for low-education female adults in employment rates changed by -4.2 percentage points (-67%).

“White High-Educ.vs. Low-Educ.Gap Δ = 2.4pp (20%) White-Black Gap for Low-Educ.Δ = -4.2pp(-67%)”

The white-Black race gap at Q1 in the chance of earning in the top household income quintile decreased by 17% between the 1978 and 1992 birth cohorts.

“White-Black Gap at Q1 (Δ = -17%)”

The weighted correlation between national-level changes in children's household income ranks and national-level changes in parental marriage rates is 0.84.

“Slope = 0.23 (0.06) Correlation = 0.84”

Between the 1978 birth cohort and the last available birth cohort, the white class gap in marriage rates at age 32 changed by 35%.

“White Class Gap (Δ = 35%) White-Black Gap at P=25 (Δ = -19%)”

In a county-level OLS regression weighted by the number of children in each county-by-race-by-class cell, a 1 percentage point increase in parental employment rate is associated with a 0.376 increase in child household income rank (SE 0.009) in the baseline specification.

“△ Parental Emp.Rate 0.376 0.399 0.308 0.293 (0.009) (0.010) (0.026) (0.026)”

When the specification in Column 2 is replicated and the sample is restricted to children whose own parents are employed, the estimates are very similar.

“Column 3 of Table II shows that we obtain very similar estimates when we replicate the specification in Column 2 and restrict the sample to children whose own parents are employed.”

The correlation between changes in white and Black children's household income ranks between the 1978 and 1992 birth cohorts at P=75 is 0.70.

“Correlation = 0.70”

Figure VIIIa shows a strong and statistically significant positive relationship between the difference in siblings' outcomes and changes in parental employment rates across cohorts in the destination county.

“Figure VIIIa shows that there is a strong and statistically significant positive relationship between the difference in siblings' outcomes and changes in parental employment rates across cohorts in the county to which they move.”

Between the 1978 and 1988 birth cohorts, the white class gap in incarceration rates changed by 32.4%.

“Δ = -11.3%Δ = 32.4%”

Between the 1978 and 1992 birth cohorts, the white-Black race gap at P=25 in parental employment rates changed by -6.4 percentage points (72%).

“White Class Gap Δ = 7.7pp (61%) White-Black Gap at P=25 Δ = -6.4pp(72%)”

The white class gap in child household income in $1,000s increased by 18% between the 1978 and 1992 birth cohorts.

“White Class Gap (Δ = 18%)”

For white children born to parents at the 25th percentile of the parental income distribution, median household income during childhood (ages 13-17) declined by $5,772 between the 1978 and 1992 birth cohorts.

“Median HH Income, Child Ages 13-17 $37,800 $32,020 -$5,772”

The pooled slope of the change in children's household income rank in adulthood on the change in parental employment rate is 0.36 (SE 0.01).

“Pooled Slope = 0.36 (0.01)”

The pooled slope of the change in children's household income rank on the change in county-level parental household income percentile at child age 27 is 0.64 (SE 0.01).

“Pooled Slope = 0.64 (0.01)”

The slope of the national-level change in children's household income rank in adulthood on the change in parental mortality rate is -1.86 (SE 0.27).

“Slope = -1.86(0.27) Correlation = -0.93”

Between the 1978 and 1988 birth cohorts, the white-Black race gap in incarceration rates at P=25 changed by -11.3%.

“Δ = -11.3%Δ = 32.4%”

For Black children born to parents at the 75th percentile of the parental income distribution, median household income during childhood (ages 13-17) increased by $1,543 between the 1978 and 1992 birth cohorts.

“Median HH Income, Child Ages 13-17 $37,260 $31,840 -$5,412 $121,700 $123,300 $1,543”

The slope of the national-level change in children's household income rank in adulthood on the change in parental marriage rate is 0.23 (SE 0.06).

“Slope = 0.23 (0.06) Correlation = 0.84”

The pooled slope of the change in children's mortality rate at ages 24-27 on the change in CZ-level parental employment rate is -0.13 (SE 0.01).

“Pooled Slope = -0.13(0.01)”

For children in the late (1987-1992) birth cohorts, the slope of the coefficient on the change in destination-county parental employment rate across move ages is -0.020 (SE 0.002).

“1987-1992 Slope = -0.020(0.002)”

Dividing children into below-median versus above-median parental income, the slope of the change in children's mean household income rank on the change in parental employment rate is 0.37 (SE 0.01).

“Pooled Slope Slope = 0.37 (0.01)”

For Black children at P=75, the slope of the change in child household income percentile (1978-1992) on baseline child household income percentile (1978 cohort) is -0.59 (SE 0.04).

“Slope = -0.59(0.04)”

The white-Black race gap in parental employment rates for low-income families decreased by 6.4 percentage points between the 1978 and 1992 birth cohorts.

“while the white-Black race gap in parental employment rates for low-income families decreased by 6.4 percentage points over the same period.”

The slope of the national-level change in children's household income rank in adulthood on the change in parental employment rate is 0.37 (SE 0.06).

“Slope = 0.37 (0.06) Correlation = 0.91”

In tax data, the white-Black gap for low-education male adults in employment rates changed by -1.3 percentage points (-18%).

“White High-Educ.vs. Low-Educ.Gap Δ = 5.3pp (67%) White-Black Gap for Low-Educ.Δ = -1.3pp(-18%)”

In tax data, the white-Black gap for low-education female adults in employment rates changed by -3.6 percentage points (-656%).

“White High-Educ.vs. Low-Educ.Gap Δ = 2.7pp (30%) White-Black Gap for Low-Educ.Δ = -3.6pp(-656%)”

Between the 1978 and 1992 birth cohorts, the white-Black race gap at P=25 in four-year college graduation rates changed by -8%.

“White Class Gap (∆ = 29%) White-Black Gap at P=25 (∆ = -8%)”

Controlling for county-level parental employment rates, the estimated change in the white-Black race gap is -4.16.

“3.37 -1.79 -1.38 -4.16”

For children who move between ages 13-17, the slope of child household income rank on the change in county-level parental employment rate in the destination county is 0.04 with a standard error of 0.05.

“Slope = 0.04 (0.05)”

The white-Black race gap at P=25 in child household income in $1,000s decreased by 28% between the 1978 and 1992 birth cohorts.

“White-Black Gap at P=25 (Δ = -28%)”

The pooled slope of the change in children's household income rank on the change in county-level parental employment rate in 2019 is 0.138 (SE 0.003).

“Pooled Slope = 0.138 (0.003)”

For white children at P=25, household income percentile at age 27 declined by 2.4 percentile points between the 1978 and 1992 cohorts.

“Household Income Percentile, Age 27 47.9 45.5 -2.4”

For siblings with a less than 4 year age gap, the slope is significantly smaller than for siblings with a 4 or more year age gap.

“Figure VIIIb replicates this analysis for siblings with a less than 4 year age gap (s( f 2 )s( f 1 ) < 4).For this group, we see a significantly smaller slope, consistent with the fact that siblings closer in age have more similar exposure to the improving community.”

For Black children at P=25, household income percentile at age 27 increased by 1.8 percentile points between the 1978 and 1992 cohorts.

“Household Income Percentile, Age 27 33.935.7 1.8”

In sibling comparisons, the implied exposure effect of growing up from birth in a community with 1 percentage point higher parental employment rates on children's household income ranks in adulthood is 0.299 (SE 0.060) for all sibling pairs.

“Implied Exposure Effect (β µ ) 0.299 0.319 0.301 0.260 (0.060) (0.092) (0.055) (0.062)”

The white class gap in children's mean household income ranks in adulthood increased by 28% between the 1978 and 1992 birth cohorts, from 11.1 to 14.1 percentiles.

“The white class gap increased by 28% between the 1978 and 1992 birth cohorts, from 11.1 to 14.1 percentiles”

The slope of the change in children's household income rank in adulthood on the change in the employment rate for same-race adults in their childhood counties is 0.21 (SE 0.03).

“Slope = 0.21 (0.03)”

The white class gap in child individual income percentile increased by 29% between the 1978 and 1992 birth cohorts.

“White Class Gap (∆ = 29%)”

Restricting to children whose own parents are employed, the pooled slope of the change in children's household income rank on the change in county-level parental employment rate at child age 27 is 0.47 (SE 0.01).

“Pooled Slope = 0.47 (0.01)”

For Black children born to parents at the 25th percentile of the parental income distribution, median household income during childhood (ages 13-17) declined by $5,412 between the 1978 and 1992 birth cohorts.

“Median HH Income, Child Ages 13-17 $37,260 $31,840 -$5,412”

For children in the early (1978-1983) birth cohorts, the slope of the coefficient on the change in destination-county parental employment rate across move ages is -0.007 (SE 0.003).

“1978-1983 Slope = -0.007(0.003)”

In the primary analysis sample, the mean household income percentile at child ages 13-17 is 58.0 for White children and 33.5 for Black children.

“Mean Household Income Percentile, Child Ages 13-17 58.0 33.5 37.7 51.3 36.9”

Under Assumption 2, the outcomes of children who move at the end of childhood to communities with increasing parental employment rates do not vary significantly across cohorts (γ1 ≈ 0 in Column 1 of Table II).

“Under Assumption 2, the relationship plotted in Figure II that the outcomes of children who move at the end of childhood to communities with increasing parental employment rates do not vary significantly across cohorts (i.e., γ 1 ≈ 0 in Column 1 of Table II).”

• AssumptionThe authors state this as an assumption: Selection occurs at the family level: differences in unobserved inputs between siblings are orthogonal to changes in parental employment rates.

In ACS data, the white high-education versus low-education gap in male adult employment rates changed by 1.9 percentage points (17%).

“White High-Educ.vs. Low-Educ.Gap Δ = 1.9pp (17%) White-Black Gap for Low-Educ.Δ = -0.3pp(-2%)”

Between the 1978 and 1992 birth cohorts, the white class gap in high school graduation rates changed by -25%.

“White Class Gap (∆ = -25%) White-Black Gap at P=25 (∆ = -41%)”

The white class gap in individual income percentile for male children increased by 28% between the 1978 and 1992 birth cohorts.

“White Class Gap (∆ = 28%)”

The white class gap in individual income percentile for female children increased by 30% between the 1978 and 1992 birth cohorts.

“White Class Gap (Δ = 30%)”

For children who move before age 8, the slope of child household income rank on the change in county-level parental employment rate in the destination county is 0.26 with a standard error of 0.04.

“Slope = 0.26 (0.04)”

The correlation between changes in white and Black children's household income ranks between the 1978 and 1992 birth cohorts at P=25 is 0.58.

“Correlation = 0.58”

The white-Black race gap at P=25 in individual income percentile for female children decreased by 120% between the 1978 and 1992 birth cohorts.

“White-Black Gap at P=25 (Δ = -120%)”

The pooled slope of the change in children's household income rank on the change in county-level parental household income percentile at child age 22 is 0.50 (SE 0.01).

“Pooled Slope = 0.50 (0.01)”

Between the 1978 birth cohort and the last available birth cohort, the white class gap in marriage rates at age 27 changed by 16%.

“White Class Gap (Δ = 16%) White-Black Gap at P=25 (Δ = -33%)”

With no controls, the estimated change in the white class gap between the 1978 and 1992 birth cohorts is 3.37.

“3.37 -1.79 -1.38 -4.16”

Controlling for county-level parental employment rates, the estimated change in the white class gap is -1.38.

“3.37 -1.79 -1.38 -4.16”

Between the 1978 and 1992 birth cohorts, the white class gap in parental employment rates changed by 7.7 percentage points (61%).

“White Class Gap Δ = 7.7pp (61%) White-Black Gap at P=25 Δ = -6.4pp(72%)”

The white-Black race gap at Q1 in the chance of earning in the bottom household income quintile decreased by 72% between the 1978 and 1992 birth cohorts.

“White-Black Gap at Q1 (Δ = -72%)”

In tax data, the white high-education versus low-education gap in female adult employment rates changed by 2.7 percentage points (30%).

“White High-Educ.vs. Low-Educ.Gap Δ = 2.7pp (30%) White-Black Gap for Low-Educ.Δ = -3.6pp(-656%)”

Among Black children born to parents in the top household income quintile, the probability of remaining in the top quintile rose from 14.5% in the 1978 birth cohort to 17.2% in the 1992 birth cohort.

“Child HH Income in Q5 3.0% 4.9% 7.2% 10.0% 14.5%”

• Also stated in: § FIGURE VII Causal Exposure Effects of Changes in Childhood Environments by Move Age and Birth Cohort · plain

  “Child HH Income in Q5 3.0% 4.4% 6.5% 10.2% 17.2%”

For Black children at P=25, a 1 percentage point increase in parental employment rate is associated with a 0.271 increase in child household income rank (SE 0.026) in the baseline specification.

“△ Parental Emp.Rate 0.271 0.256 0.114 0.084 (0.026) (0.026) (0.021) (0.019)”

The white class gap in parental employment rates increased by 7.7 percentage points between the 1978 and 1992 birth cohorts.

“The white class gap in parental employment rates increased by 7.7 percentage points between the 1978 and 1992 birth cohorts”

With no controls, the estimated change in the white-Black race gap between the 1978 and 1992 birth cohorts is -1.79.

“3.37 -1.79 -1.38 -4.16”

In ACS data, the white high-education versus low-education gap in female adult employment rates changed by 2.4 percentage points (20%).

“White High-Educ.vs. Low-Educ.Gap Δ = 2.4pp (20%) White-Black Gap for Low-Educ.Δ = -4.2pp(-67%)”

The white class gap in the chance of earning in the bottom household income quintile increased by 31% between the 1978 and 1992 birth cohorts.

“White Class Gap (Δ = 31%)”

Between the 1978 and 1992 birth cohorts, the white-Black race gap at P=25 in high school graduation rates changed by -41%.

“White Class Gap (∆ = -25%) White-Black Gap at P=25 (∆ = -41%)”

In tax data, the white high-education versus low-education gap in male adult employment rates changed by 5.3 percentage points (67%).

“White High-Educ.vs. Low-Educ.Gap Δ = 5.3pp (67%) White-Black Gap for Low-Educ.Δ = -1.3pp(-18%)”

The pooled slope of the change in children's household income rank on the change in CZ-level parental mortality rate between child ages 18-27 is -1.70 (SE 0.10).

“Pooled Slope = -1.70(0.10)”

The white-Black race gap at P=25 in individual income percentile for male children decreased by 29% between the 1978 and 1992 birth cohorts.

“White-Black Gap at P=25 (∆ = -29%)”

For white children born to parents at the 75th percentile of the parental income distribution, median household income during childhood (ages 13-17) increased by $2,430 between the 1978 and 1992 birth cohorts.

“Median HH Income, Child Ages 13-17 $37,800 $32,020 -$5,772 $122,100 $124,500 $2,430”

In the specification controlling for adult county-by-parental income percentile-by-race-by-cohort fixed effects, the childhood exposure effect is βμ = 0.445.

“We continue to find a substantial childhood exposure effect in this specification ( βµ = 0.445), indicating that differential exposure to labor demand shocks is unlikely to drive our baseline results.”

The pooled slope of the change in children's household income rank on the change in county-level parental employment rate in 2012 is 0.21 (SE 0.01).

“Pooled Slope = 0.21 (0.01)”

The slope of the between-group relationship between national-level changes in children's household income ranks and changes in parental employment rates is 0.37, nearly identical to the within-group slope in Figure Va.

“The slope of the between-group relationship in Appendix Figure A.27 is 0.37, nearly identical to the within-group slope in Figure Va.”

Changes in outcomes across cohorts for white and Black children raised in low-income families have a strong positive correlation of 0.58 across areas.

“The changes in outcomes across cohorts for white and Black children raised in low-income families have a strong positive correlation of 0.58 across areas”

The relationship between children's and parents' income ranks is flatter than in Chetty et al. (2014a) and Chetty et al. (2020) because children's income is measured at age 27, compared to ages 29-30 and 31-37 respectively.

“The relationship between children's and parents' income ranks is flatter than inChetty et al. (2014a) andChetty et al. (2020) because we measure children's income at age 27, compared to ages 29-30 inChetty et al. (2014a) and ages 31-37 inChetty et al. (2020).”

• cites Chetty et al. (2014a)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Chetty et al. (2020) sourceNot checked yet

The outcomes of white children growing up in low-income families are much more strongly related to employment rates of low-income white parents than high-income white parents.

“Furthermore, the outcomes of white children growing up in low-income families are much more strongly related to employment rates of lowincome white parents than high-income white parents”

There are similarly strong correlations between national-level changes in children's outcomes and changes in parental marriage rates and parental mortality rates for all race and class groups.

“We find similarly strong correlations between nationallevel changes in children's outcomes and changes in parental marriage rates and parental mortality rates for all race and class groups”

The relationship between changes in children's outcomes and changes in community-level parental employment rates is not driven by children's own parents' employment status.

“The relationship between changes in children's outcomes and changes in community-level parental employment rates is not driven by children's own parents' employment status.”

• Also stated in: § Year Trend · plain

  “When we replicate FigureVaon the subsample of children whose parents are employed when they are age 27, the relationship between changes in children's outcomes and changes in parental employment rates persists”

While levels of rank-rank slopes differ when children's incomes are measured at earlier ages, trends across cohorts are similar.

“While the levels of the rank-rank slopes differ when children's incomes are measured at earlier ages, the trends across cohorts are similar, as shown in Appendix FigureA.3.”

Changes in the employment rates of low-income Black parents in the same county are unrelated to changes in white children's outcomes, conditional on the change for white parents.

“In contrast, changes in the employment rates of low-income Black parents in the same county are unrelated to changes in white children's outcomes (conditional on the change for white parents).”

Economic mobility increased sharply in the Southeast and the industrial Midwest between the 1978 and 1992 birth cohorts.

“For example, economic mobility increased sharply in the Southeast and the industrial Midwest between the 1978 and 1992 birth cohorts”

Most of the variation in parental employment rates is across cohorts rather than calendar years.

“consistent with the fact that most of the variation in parental employment rates is across cohorts rather than calendar years, as shown in Appendix FigureA.26.”

• cites Appendix Figure A.26Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

The estimate of γ1 ≈ 0 in Column 1 of Table II shows that the outcomes of children who move late in childhood to areas with increasing employment rates do not change across cohorts.

“The estimate of γ 1 ≈ 0 in Column 1 of Table II (and the corresponding non-parametric evidence in Figure VIIc) shows that the outcomes of children who move late in childhood to areas with increasing employment rates do not change across cohorts.”

Children who move at young ages are exposed to higher community-level employment rates during childhood in later cohorts.

“Similar to the reduced-form patterns in Figure VII, children who move at young ages are exposed to higher community-level employment rates during childhood in later cohorts.”

The correlation between national-level changes in children's household income ranks and changes in parental employment rates for all race and class groups is 0.91.

“The correlation between these variables is 0.91.”

The white class gap in the share of children taking the SAT/ACT increased by 12.1 percentage points between the 1980 and 1991 birth cohorts.

“The white class gap in the share of children taking the SAT/ACT increased by 12.1 percentage points between the 1980 and 1991 birth cohorts”

White parents' employment rates have little impact on Black children's outcomes in counties where Black children have below-median exposure to white children.

“In contrast, white parents' employment rates have little impact on Black children's outcomes in counties where Black children have below-median exposure to white children.”

Conditioning on X_i using more flexible methods such as non-parametric or propensity score reweighting yields very similar conclusions.

“Conditioning on X i using more flexible methods such as non-parametric or propensity score reweighting yields very similar conclusions.”

The average level of outcomes of children in the 1978 cohort is uncorrelated with changes in parental employment rates from the 1978-1992 cohorts across communities, controlling for baseline parental employment rates.

“This result for movers is consistent with the fact that the average level of outcomes of children in the 1978 cohort ȳcpr,s=1978 is uncorrelated with changes in parental employment rates from the 1978-1992 cohorts ∆ ēcpr across communities, controlling for ēcpr,s=1978(Appendix Figure A.36).”

• cites Appendix Figure A.36Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

There is a positive relationship between changes in parental employment rates and children's outcomes for Black children raised in high-income families.

“We find a positive relationship between changes in parental employment rates and children's outcomes for Black children raised in high-income families as well, but estimates are noisier due to small sample sizes in many counties(Appendix Figure A.20).”

• cites Appendix Figure A.20Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Measures of adults' employment rates by race and education constructed from publicly available cross-sectional datasets such as the ACS show qualitatively similar changes in the white class and white-Black race gaps, although magnitudes are attenuated relative to longitudinal data.

“Measures of adults' employment rates by race and education (an alternative proxy for class) constructed from publicly available cross-sectional datasets such as the ACS also show qualitatively similar changes in the white class and white-Black race gaps, although the magnitudes of the changes are attenuated relative to what we observe in our longitudinal data for reasons discussed in Appendix B.”

Black children's outcomes are much more strongly associated with changes in Black parents' employment rates in the counties in which they grow up than with white parents' employment rates.

“Here, we find the opposite pattern: Black children's outcomes are much more strongly associated with changes in Black parents' employment rates in the counties in which they grow up than with white parents' employment”

Changes in the parental employment rates of white children in a given county are positively associated with changes in white children's outcomes, with a coefficient of βw = 0.29.

“Changes in the parental employment rates of white children in a given county are positively associated with changes in white children's outcomes, with a coefficient of βw = 0.29”

For white children born to families in the top household income quintile, the chances of falling to the bottom quintile did not change significantly.

“For white children born to families in the top household income quintile, the chances of remaining in the top quintile increased while chances of falling to the bottom quintile did not change significantly.”

Labor demand shocks and selection correlated with parental employment drive very little of the change in children's outcomes across cohorts.

“Fortunately, the orange series in FigureVIIcsuggests that labor demand shocks and selection correlated with parental employment drive very little of the change in children's outcomes across cohorts, explaining why we obtain similar estimates using this approach relative to our baseline approach.”

Black children born in 1992 in counties with the highest levels of upward mobility for Black children still have poorer outcomes in adulthood on average than white children born in counties with the lowest levels of upward mobility for white children.

“Black children born in 1992 in counties with the highest levels of upward mobility for Black children still have poorer outcomes in adulthood on average than white children born in counties with the lowest levels of upward mobility for white children.”

Changes in low-income white parents' employment rates are highly predictive of changes in low-income Black children's outcomes in counties where Black children are more exposed to white peers, either in schools or neighborhoods.

“Consistent with prior evidence of heterogeneous social interaction based on white shares, changes in low-income white parents' employment rates are highly predictive of changes in low-income Black children's outcomes in counties where Black children are more exposed to white peers, either in schools or neighborhoods.”

For white children born to families in the top household income quintile, the chances of remaining in the top quintile increased.

“For white children born to families in the top household income quintile, the chances of remaining in the top quintile increased while chances of falling to the bottom quintile did not change significantly.”

Despite gains, there are still vast and widespread white-Black race gaps in economic mobility for children born to low-income families even in the 1992 birth cohort.

“Despite these gains, there are still vast and widespread white-Black race gaps in economic mobility for children born to low-income families even in the 1992 birth cohort”

The slope of the relationship between moving early to counties with increased parental employment rates and children's outcomes is 0.26 for the 1992 birth cohort.

“The slope of this relationship (0.26) can be interpreted as the impact of a 1 percentage point increase in ∆ ēdpr on children's outcomes for the 1992 birth cohort”

The white-Black race gap in reaching the top household income quintile conditional on being born to families in the bottom quintile fell by only 1.9 percentage points (17%) between the 1978 and 1992 birth cohorts.

“The white-Black race gap using this measure fell by only 1.9 percentage points (17%) between the 1978 and 1992 birth cohorts”

Children in the 1992 cohort who moved early in childhood (before age 8) to counties where parental employment rates increased significantly in their own race-by-class group over the prior 15 cohorts have significantly better outcomes in adulthood.

“Figure VIIa shows that children in the 1992 cohort who moved early in childhood to counties where parental employment rates increased significantly in their own race-by-class group over the prior 15 cohorts have significantly better outcomes in adulthood.”

The areas in which outcomes of children born to low-income Black families improved the most tend to be areas in which outcomes of children born to low-income white families deteriorated the least.

“The areas in which outcomes of children born to low-income Black families improved the most tend to be areas in which outcomes of children born to low-income white families deteriorated the least.”

Focusing on calendar year can mask cohort-level trends because each birth cohort is a small share of the working age population at any point in time.

“Focusing on calendar year can mask cohort-level trends because each birth cohort is a small share of the working age population at any point in time.”

The divergence in economic mobility must be driven by factors that impact race and class groups differently within a given neighborhood.

“The divergence in economic mobility must therefore be driven by factors that impact race and class groups differently within a given neighborhood.”

Past work examining white-Black income gaps such as Bayer and Charles (2018) did not find these trends because it disaggregated the data by calendar year rather than by birth cohort.

“These trends were not evident in past work examining white-Black income gaps such as Bayer and Charles (2018) because that work disaggregated the data by calendar year rather than by birth cohort.”

• cites Bayer and Charles (2018) sourceNot checked — no full text of the source

The white-Black race gap among low-income families narrowed primarily because of changes in children's chances of escaping poverty rather than their chances of reaching the upper class.

“The white-Black race gap among low-income families narrowed primarily because of changes in children's chances of escaping poverty rather than their chances of reaching the upper class.”

Community-level changes-as proxied by parental employment rates or other outcomes in the parental generation-provide a unified explanation (in a predictive sense) for the divergence in outcomes by race and class.

“communitylevel changes-as proxied by parental employment rates or other outcomes in the parental generationprovide a unified explanation (in a predictive sense) for the divergence in outcomes by race and class.”

• AssumptionThe authors state this as an assumption: Parental employment rates proxy for childhood environments.

Under the identification assumption, changes in communities across cohorts as proxied by parental employment rates lead to an increase in children's earnings through a causal exposure effect.

“Under our identification assumption, these results imply that changes in communities across cohorts as proxied by parental employment rates lead to an increase in children's earnings through a causal exposure effect.”

• AssumptionThe authors state this as an assumption: Selection effects across cohorts do not differ by the age at which children move (constant selection by age).
• AssumptionThe authors state this as an assumption: Parental employment rates proxy for childhood environments.

The relationship between children's outcomes and parental employment rates is mediated by social interaction.

“suggesting that the relationship between children's outcomes and parental employment rates is mediated by social interaction.”

Divergent trends in economic mobility were caused by differential changes in childhood environments, as proxied by parental employment rates, within local communities defined by race, class, and childhood county.

“Using a quasi-experimental design, we show that the divergent trends in economic mobility were caused by differential changes in childhood environments, as proxied by parental employment rates, within local communities defined by race, class, and childhood county.”

• AssumptionThe authors state this as an assumption: Parental employment rates proxy for childhood environments.

A parsimonious theory-that children's outcomes mimic those of the parents in their social communities, as in Borjas (1992)-explains the divergent trends in economic mobility by race and class in the United States in recent decades.

“Combining these results, we conclude that a parsimonious theory-that children's outcomes mimic those of the parents in their social communities, as in Borjas (1992)-explains the divergent trends in eco-nomic mobility by race and class in the United States in recent decades.”

• cites Borjas (1992) sourceNot checked — no full text of the source

The cohort-specificity of impacts points in favor of social interaction mechanisms.

“Insofar as economic resources are unlikely to vary so sharply across adjacent cohorts, this cohort-specificity of impacts points in favor of social interaction mechanisms.”

The white-Black race gap for children born to low-income families narrowed.

“As a result, the white-Black race gap for children born to low-income families narrowed.”

Results are similar when examining alternative measures of income such as individual earnings or household income measured in dollars rather than percentile ranks.

“We find similar patterns of growing white class gaps and shrinking white-Black race gaps when examining alternative measures of income such as individual earnings or household income measured in dollars rather than percentile ranks.”

Increases in parental employment rates translate to much smaller changes in children's outcomes if children arrive in the destination county late in childhood.

“This relationship is much flatter across cohorts than the corresponding series for young movers, implying that increases in parental employment rates translate to much smaller changes in children's outcomes if they arrive in the destination county late in childhood.”

Changes in community-level parental employment rates can explain the divergent trends in both monetary and non-monetary outcomes documented above.

“showing that changes in community-level parental employment rates can explain the divergent trends in both monetary and non-monetary outcomes documented above.”

Under the identification assumption in Equation (9), changes in childhood environments drive much of the correlation between changes in parental employment rates and changes in children's outcomes documented in Figure V.

“Under the identification assumption in Equation ( 9), this result implies that changes in childhood environments drive much of the correlation between changes in parental employment rates and changes in children's outcomes documented in Figure V.”

• AssumptionThe authors state this as an assumption: The covariance between changes in unobserved family inputs θ i and changes in parental employment rates ēdpr does not vary with the child's age of move m.

The county-by-race measures of employment used in Appendix Figure A.17 cannot explain the divergent trends across race and class groups because they do not vary across class.

“cannot explain the divergent trends across race and class groups because they do not vary across class.”

• AssumptionThe authors state this as an assumption: Any measurement error or bias in children's and parents' incomes does not change across cohorts in ways that differ by race or class.
• AssumptionThe authors state this as an assumption: The composition of the sample does not change across cohorts in ways that differ by race or class.

The relationship between Black children's and parents' income ranks shifted upward rather than steepened.

“leading to an upward shift rather than a steepening of the relationship between children's and parents' income ranks.”

The absence of significant changes in outcomes across cohorts for late-childhood movers directly shows that changes in childhood environments are responsible for the changes observed in children's outcomes.

“The absence of such changes directly shows that changes in childhood environments are responsible for the changes we observe in children's outcomes.”

The sharp decay in the impacts of parental employment rates across cohorts points in favor of a social interaction mechanism.

“The sharp decay in the impacts of parental employment rates across cohorts thus points in favor of a social interaction mechanism.”

The impacts of changes in childhood environments on children's outcomes are proportional to the number of years of exposure they have to that environment, with roughly constant dosage effects throughout childhood.

“The impacts of changes in childhood environments on children's outcomes are thus proportional to the number of years of exposure they have to that environment, with roughly constant dosage effects throughout childhood.”

When the sample is restricted to children who spent their entire childhood in a single county, the estimates are similar to baseline.

“We obtain similar estimates when we restrict the sample to children who spent their entire childhood in a single county.”

Children raised in low-income families are increasingly likely to grow up with a single parent compared to children raised in high-income families.

“children raised in low-income families are increasingly likely to grow up with a single parent compared to children raised in high-income families (e.g., Lundberg, 2017;Kearney, 2022).”

• cites Lundberg, 2017 sourceNot checked yet

  Checked before — its earlier verdict is held back until a new check runs.

• cites Kearney, 2022 sourceNot checked yet

  Checked before — its earlier verdict is held back until a new check runs.

Results are insensitive to how children's and parents' incomes are measured or how the sample is constructed.

“We conduct several robustness checks to evaluate these concerns and find that that our results are insensitive to how we measure children's and parents' incomes or how we construct our sample (Appendix Figure A.9).”

• AssumptionThe authors state this as an assumption: Any measurement error or bias in children's and parents' incomes does not change across cohorts in ways that differ by race or class.
• AssumptionThe authors state this as an assumption: The composition of the sample does not change across cohorts in ways that differ by race or class.

Social interaction mediates changes in economic mobility.

“Children's outcomes are more strongly related to the parental employment rates of peers they are more likely to interact with, suggesting that social interaction mediates changes in economic mobility.”

The changes in mean SAT/ACT scores are likely attenuated by the changing selection of test takers during our sample period.

“The changes in mean SAT/ACT scores are likely attenuated by the changing selection of test takers during our sample period.”

Changes in parental employment rates largely capture work patterns for the parents of a given cohort rather than year-specific labor market shocks.

“These findings show that changes in parental employment rates largely capture work patterns for the parents of a given cohort rather than year-specific labor market shocks.”

The white-Black race gap for children from high-income families remained essentially unchanged.

“The white-Black race gap for children from high-income families thus remained essentially unchanged.”

In a linear model with homogeneous treatment effects of community-level factors on children's outcomes, the structural coefficient on changes in community-level factors is invariant to aggregation.

“Formally, in a linear model with homogeneous treatment effects of community-level factors on children's outcomes, the structural coefficient on changes in community-level factors is invariant to aggregation.”

The similarity of results using household income and individual-level earnings shows that the trends are not driven by changes in the number of adults in the household.

“The fact that results are similar when using household income and individual-level earnings shows that the trends are not driven by changes in the number of adults in the household.”

Sibling comparisons net out unobservables that are fixed within families, allowing the identification assumption underlying the estimator to be weakened.

“Sibling comparisons net out unobservables that are fixed within families, allowing us to weaken the identification assumption underlying our estimator.”

Many mechanisms that operate at a local level will aggregate to the county level.

“However, many mechanisms that operate at a local level will aggregate to the county level.”

The divergent trends in economic mobility must be driven by changes that affect race and class groups differently within neighborhoods.

“The preceding results indicate that the divergent trends in economic mobility must be driven by changes that affect race and class groups differently within neighborhoods.”

• AssumptionThe authors state this as an assumption: Any measurement error or bias in children's and parents' incomes does not change across cohorts in ways that differ by race or class.
• AssumptionThe authors state this as an assumption: The composition of the sample does not change across cohorts in ways that differ by race or class.

Trends in children's earnings diverge sharply by race and class even among children who grow up in the same neighborhood and who share the same observable family characteristics.

“In short, trends in children's earnings diverge sharply by race and class even among children who grow up in the same neighborhood and who share the same observable family characteristics.”

• AssumptionThe authors state this as an assumption: Any measurement error or bias in children's and parents' incomes does not change across cohorts in ways that differ by race or class.
• AssumptionThe authors state this as an assumption: The composition of the sample does not change across cohorts in ways that differ by race or class.

The link between changes in children's outcomes and changes in parental employment rates cannot be a mechanical consequence of changes in labor market opportunities, because educational outcomes are measured before children enter the labor market.

“Since the educational outcomes are measured before children enter the labor market, the link between changes in children's outcomes and changes in parental employment rates cannot be a mechanical consequence of changes in labor market opportunities.”

The changes in children's outcomes at the bottom of the income distribution are driven largely by changes in the fraction of children who were employed at age 27.

“The changes in children's outcomes at the bottom of the income distribution are driven largely by changes in the fraction of children who were employed at age 27 (Figure IIa and Appendix Figure A.6).”

Observable family characteristics explain a small share of the growing white class gap in intergenerational mobility.

“Hence, observable family characteristics explain a small share of the growing white class gap in intergenerational mobility.”

Differential changes in the parental characteristics observed in the data do not explain the baseline findings.

“These findings show that differential changes in the parental characteristics that we observe in our data do not explain our baseline findings.”

Estimates of changes in mobility in smaller geographies are very noisy; reliability estimates range from 0.614-0.784 at the county level compared to 0.024-0.105 at the Census tract level.

“Reliability estimates-the share of the variance in the estimates due to signal rather than sampling noise-range from 0.614-0.784across subgroups for changes in children's mean income ranks across cohorts at the county level, compared to just 0.024-0.105at the Census tract level (Appendix Table A.21).”

The paper's baseline estimates of β μ rely on the 'constant selection by age' identification assumption in Equation (9).

“Our baseline estimates of β µ rely on the”

• AssumptionThe authors state this as an assumption: The covariance between changes in unobserved family inputs θ i and changes in parental employment rates ēdpr does not vary with the child's age of move m.

Under a linear approximation, a county-level analysis offers a statistically viable approach to identify community-level relationships of interest even when the mechanisms operate at narrower geographies.

“Hence, under a linear approximation, a county-level analysis offers a statistically viable approach to identify community-level relationships of interest even when the mechanisms operate at narrower geographies.”

Defining class by parental income during childhood reveals important trends in parental employment beyond what can be captured by defining class by only parental education.

“These results show that defining class by parental income during childhood reveals important trends in parental employment, beyond what we can capture when we define class by only parental education.”

The stronger influence of own-group parental employment rates on children's outcomes is consistent with the high degree of homophily by race in social interactions.

“The stronger influence of own-group parental employment rates on children's outcomes is consistent with the high degree of homophily by race in social interactions.”

Most of the changes in the white class gap and white-Black race gap can be explained by the sharp fall in employment rates for low-income white parents relative to low-income Black and high-income white parents during the period studied.

“This result implies that most of the changes in the white class gap and white-Black race gap can be explained by the sharp fall in employment rates for low-income white parents relative to low-income Black and high-income white parents during the period we study (Appendix Figure A.22).”

Economic resources are unlikely to vary sharply across adjacent cohorts: having more employed parents in the grade before or after one's grade would presumably contribute similarly to greater resources in the community.

“Economic resources are unlikely to vary sharply across adjacent cohorts: having more employed parents in the grade before or after one's grade would presumably contribute similarly to greater resources in the community.”

These divergent trends in economic mobility by race and class were driven by differential changes in the social environments in which children grew up.

“These divergent trends in economic mobility by race and class were driven by differential changes in the social environments in which children grew up.”

Students on the margin of taking the SAT/ACT tend to have lower scores than students who always take these tests.

“Students on the margin of taking the SAT/ACT tend to have lower scores than students who always take these tests (Dynarski, 1987;Dynarski and Gleason, 1993;Clark, Rothstein and Schanzenbach, 2009).”

• cites Dynarski, 1987 sourceNot checked — no full text of the source
• cites Dynarski and Gleason, 1993 sourceNot checked — no full text of the source
• cites Clark, Rothstein and Schanzenbach, 2009Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Children's outcomes are driven by changes in their broader community rather than changes in their own parents' employment rates.

“This result shows that children's outcomes are driven by changes in their broader community rather than changes in their own parents' employment rates, consistent with the observational analysis in Appendix Figure A.24e.”

Under the assumption that children who move to better areas at younger ages are weakly positively selected, the finding that children's outcomes in the 1978 cohort do not vary with age at move to a destination with increasing parental employment rates implies that the baseline level of destination place effects is uncorrelated with the change in destination parental employment rates.

“Under the assumption that children who move to better areas at younger ages are weakly positively selectedi.e., Cov( θdpr,m,s=1978 -θdpr,m+1,s=1978 , ∆ ēdpr ) ≥ 0 ∀m -this finding implies that the baseline level of destination place effects is uncorrelated with the change in destination parental employment rates, i.e., Cov(µ d pr,s=1978 , ∆ ēdpr ) = 0”

• AssumptionThe authors state this as an assumption: Children who move to better areas at younger ages are weakly positively selected.

Under weak positive selection assumptions, the siblings estimator β f can be expressed as β μ rescaled by a constant.

“Under the weak positive selection assumptions described above, we show in Appendix F that the siblings estimator β f can be expressed as β µ rescaled by a constant:”

• AssumptionThe authors state this as an assumption: Selection occurs at the family level: differences in unobserved inputs between siblings are orthogonal to changes in parental employment rates.
• AssumptionThe authors state this as an assumption: Children who move to better areas at younger ages are weakly positively selected.

The difference in reduced-form slopes between the two groups is driven by the fact that siblings' outcomes respond to changes in parental employment rates in direct proportion to the sibling age gap.

“The difference in reduced-form slopes between the two groups is driven by the fact that siblings' outcomes respond to changes in parental employment rates in direct proportion to the sibling age gap, as shown by Equation (E.4).”

Level differences in neighborhood effects do not drive the baseline estimates.

“This finding suggests that level differences in neighborhood effects do not drive our baseline estimates, consistent with the evidence in Figure VIIc showing that there is little difference in outcomes for children in the 1978 cohort who move at younger versus older ages to places where parental employment rates subsequently improve.”

Differential exposure to labor demand shocks is unlikely to drive the baseline results.

“We continue to find a substantial childhood exposure effect in this specification ( βµ = 0.445), indicating that differential exposure to labor demand shocks is unlikely to drive our baseline results.”

β f provides a lower bound for β μ assuming only weak positive selection.

“β f Provides a Lower Bound for β µ Assuming Only Weak Positive Selection.”

• AssumptionThe authors state this as an assumption: Selection is weakly positive (Cov(∆ θdprm, ∆ ēdpr) ≥ 0 for all m).

The smaller slope for siblings closer in age is consistent with siblings closer in age having more similar exposure to the improving community.

“Figure VIIIb replicates this analysis for siblings with a less than 4 year age gap (s( f 2 )s( f 1 ) < 4).For this group, we see a significantly smaller slope, consistent with the fact that siblings closer in age have more similar exposure to the improving community.”

The effect of a 1 percentage point increase in parental employment rates from the 1978-1992 cohorts on the gap in siblings' outcomes is proportional to the effect of being exposed from birth to a community with 1 percentage point higher parent employment rates (β μ ) times the gap in ages between the two siblings (g( f )).

“The effect of a 1 percentage point increase in parental employment rates from the 1978-1992 cohorts on the gap in siblings' outcomes is proportional to the effect of being exposed from birth to a community with 1 percentage point higher parent employment rates (β µ ) times the gap in ages between the two siblings (g( f )).”

In practice, children who move at an earlier age in childhood are more likely to stay in the destination as adults.

“As discussed above, Sprung-Keyser and Porter (2023) show that in practice, children who move at an earlier age in childhood are more likely to stay in the destination as adults.”

• cites Sprung-Keyser and Porter (2023)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

The independence assumption that (g( f ), m( f 2 ), s( f 1 )) is balanced by destination is supported by Columns 1 and 2 of Appendix Table A.26, which show that the differences in β f by g( f ) are consistent with our model, such that the rescaled estimates βμ are nearly identical.

“This assumption is supported by Columns 1 and 2 of Appendix Table A.26, which show that the differences in β f by g( f ) are consistent with our model, such that the rescaled estimates βµ are nearly identical.”

• cites Appendix Table A.26Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• AssumptionThe authors state this as an assumption: (g( f ), m( f 2 ), s( f 1 )) is balanced by destination.

The unified explanation for differential trends by race is consistent with the sociological thesis of racial invariance.

“This unified explanation for differential trends by race is consistent with the sociological thesis of racial invariance-the idea that the ultimate causes of economic and social outcomes are the same for white and Black individuals.”

• cites Sampson and Wilson (1995) sourceNot checked — no full text of the source
• cites Sampson, Wilson and Katz (2018) sourceNot checked — no full text of the source

Accounting for differences in mortality rates is important when calculating employment rates.

“Accounting for Mortality: Finally, we show the additional importance of accounting for differences in mortality rates when calculating employment rates.”

Our findings are not driven by time-varying confounds from changes in own-family income and employment.

“supporting the view that our findings are not driven by such time-varying confounds.”

Controlling for changes in own-family income and employment has little impact on our results.

“In practice, we find that controlling for changes in own-family income and employment has little impact on our results (Column 3 of TableII), supporting the view that our findings are not driven by such time-varying confounds.”

• cites Column 3 of Table IINot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

The baseline estimate of βμ is 0.339, indicating that growing up from birth in a community with a 1 percentage point higher parental employment rate leads to an increase in children's mean household income ranks at age 27 by 0.339 ranks.

“Column 1 of Table II reports our baseline estimate of βµ = 0.339, indicating that growing up from birth in a community with a 1 percentage point higher parental employment rate leads to an increase in children's mean household income ranks at age 27 by 0.339 ranks.”

• AssumptionThe authors state this as an assumption: Assumption 1 holds: under certain regularity conditions, βµ identifies the causal effect of spending one's whole childhood in a destination county where parental employment rates increased by 1 percentage point.

Measuring parental employment rates in a fixed calendar year for all cohorts yields similar results.

“We find similar results when we measure parental employment rates in a fixed calendar year for all cohorts (Appendix Figure A.31), consistent with the fact that most of the variation in parental employment rates is across cohorts rather than calendar years, as shown in Appendix FigureA.26.”

• cites Appendix Figure A.31Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Including origin-by-parent income percentile-by-race-by-cohort-by age at move fixed effects eliminates variation arising from differences in origin quality under the additive structural model in Equation (3).

“In our empirical analysis, we include origin-by-parent income percentile-by-race-by-cohort-by age at move fixed effects, which eliminates variation arising from differences in origin quality under the additive structural model in Equation (3).”

Community-level changes in the parents' generation, measurable by a variety of parental outcomes, are strongly correlated with children's outcomes in adulthood.

“The broader point is that community-level changes in the parents' generation (which can be measured using a variety of parental outcomes) are strongly correlated with children's outcomes in adulthood.”

Differences in the probability of staying in a destination labor market by age at move do not drive the findings.

“compare the outcomes of children who live in the same destination county as adults but move to that county at different ages during childhood, showing that differences in the probability of staying in a destination labor market by age at move (as documented by Sprung-Keyser and Porter (2023)) do not drive our findings.”

• cites Sprung-Keyser and Porter (2023)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Under Assumption 1, the estimate of βμ implies that 90% of the cross-sectional relationship between changes in parental employment rates and changes in children's outcomes is driven by changes in causal childhood exposure effects.

“Hence, under Assumption 1, this estimate of β µ implies that 90% of the cross-sectional relationship between changes in parental employment rates and changes in children's outcomes is driven by changes in causal childhood exposure effects.”

• AssumptionThe authors state this as an assumption: Assumption 1 holds: under certain regularity conditions, βµ identifies the causal effect of spending one's whole childhood in a destination county where parental employment rates increased by 1 percentage point.

The similarity of the rescaled βμ estimates across sibling age gap samples serves as an overidentification test of the model, showing that the magnitude of the difference in slopes plotted in Figure VIIIa and Figure VIIIb matches what one would predict based on differences.

“The similarity of these estimates serves as an overidentification test of our model, showing that the magnitude of the difference in slopes plotted in Figure VIIIa and Figure VIIIb matches what one would predict based on differences”

The relationship between children's outcomes and parents' employment rates is not driven by correlated labor demand shocks.

“These cohort-specific patterns also provide further evidence that the relationship between children's outcomes and parents' employment rates is not driven by correlated labor demand shocks, since labor demand fluctuations are unlikely to covary at a high frequency between children and parents who have children in the same birth cohort.”

Changes in parental employment rates are just one of many community-level factors that predict changes in children's outcomes.

“These findings show that changes in parental employment rates are just one of many community-level factors that predict changes in children's outcomes.”

Baseline level differences in causal effects across destinations do not confound the estimates.

“compare the outcomes of children from different cohorts who move from the same origin to the same destination at the same age, showing that baseline level differences in causal effects µ d pr,s=1978 across destinations do not confound our estimates”

Estimating Equation (1) using OLS provides a consistent estimate of the average change in the white class gap integrating over the distribution of X_i under the parametric assumptions made in Equation (1).

“Estimating Equation (1) using OLS provides a consistent estimate of the average change in the white class gap integrating over the distribution of X i under the parametric assumptions made in Equation (1).”

• AssumptionThe authors state this as an assumption: The parametric assumptions made in Equation (1) hold.

The positive correlation between changes in outcomes for low-income white and Black children across areas indicates that gains for low-income Black families did not come directly at the expense of their white counterparts in the same areas.

“This positive correlation indicates that the gains for children born to low-income Black families did not come directly at the expense of their white counterparts in the same areas.”

Children whose parents have below-median socioeconomic status are 17% less likely to befriend children from above-median-SES families than they would if friendships were made uniformly by class.

“Homophily by class is weaker than by race but also significant: data on friendships from Facebook show that children whose parents have below-median socioeconomic status are 17% less likely to befriend children from above-median-SES families than they would if friendships were made uniformly by class(Chetty et al., 2022).”

• cites Chetty et al., 2022 sourceNot checked yet

  Checked before — its earlier verdict is held back until a new check runs.

National statistics exhibit a shrinking white-Black gap for the employment-population ratio over this period.

“For example, national statistics exhibit a shrinking white-Black gap for the employment-population ratio over this period(Bureau of Labor Statistics, 2005-2019).”

• cites Bureau of Labor Statistics, 2005-2019Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

The white-Black race gap shrank primarily because of changes in children's chances of rising out of poverty rather than their breaking into the upper class.

“The white-Black race gap shrank primarily because of changes in children's chances of rising out of poverty rather than their breaking into the upper class.”