Where Is the Land of Opportunity? The Geography of Intergenerational Mobility in the United States

• p. 27 V.C

  have ¯r25,c < 37.4

• p. 29 V.C Relative Mobility

  differs by more than 40.2 percentiles

• p. 29 V.C Relative Mobility

  less than 23.5 percentiles

• p. 89 Figure VI legends

  26.0 - 37.3

Only the authors can say what is right here.

Judge's note: The p.29 text describes the deciles of the Figure VIb map, but the Panel B legend (p.89) gives 0.404 and 0.240, i.e. 40.4 and 24.0 percentiles. Rounding of displayed bin edges could account for 37.4 vs 37.3 in Panel A, but not for the Panel B differences. The top cut-off of 52.0 does match. The paper gives no explanation, such as a different CZ set.

Second check: another model agreed

• p. 83 Online Appendix Table VIII, Local Labor Market

  Growth in Chinese Imports 1990-2000 (Autor and Dorn 2013)

• p. 65 References

  Autor, David H., David Dorn, and Gordon H. Hanson. 2013.

Judge's note: Image p83 shows the row label. The reference list (p65) gives 'Autor, David H., David Dorn, and Gordon H. Hanson. 2013', and the text says 'Autor, Dorn and Hanson'. No two-author 2013 Autor-Dorn entry exists in the reference list.

Second check: another model agreed

• p. 3 Introduction, paragraph 3

  between 1996 and 2000, when the children are between the ages of 15 and 20.

• p. 20 Section IV.B, Lifecycle bias

  While we cannot measure parent income before age 14 for children in our core sample

• p. 3 Introduction

  when the children are between the ages of 15 and 20

• p. 20 Section IV.B

  we cannot measure parent income before age 14

Only the authors can say what is right here.

Judge's note: Seen on p3 and p20. In 1996 the 1982 cohort is 14 (age by calendar year), and Section IV.B itself names 14 as the earliest age. So the introduction's lower bound of 15 conflicts with the paper's own statement. The ages are approximate, so this is minor.

Second check: another model agreed

• p. 4 Introduction, CZ paragraph

  We assign children to commuting zones based on where they lived at age 16

• p. 24 Section V.A

  we therefore typically measure location when children were approximately 15 years old

Judge's note: p4 against p24. Location is the ZIP code in the first year the child is claimed, mostly 1996. That gives ages 14-16 for the 1980-82 cohorts, which Section V.A summarises as about 15. Age 16 in the introduction fits only the 1980 cohort. This is a loose summary rather than a methods error.

Second check: another model agreed

• p. 6 Introduction, paragraph on multivariable regression

  the five key factors described above generally remain statistically significant predictors of both relative and absolute upward mobility, even in specifications with state fixed effects

• p. 75 Table VI, columns 1, 2 and 4

  0.006

• p. 43 Section VI.G, text on Column 4

  but the other factors are not statistically significant

• p. 75 Table VI col 4

  0.154

Only the authors can say what is right here.

Judge's note: Read on the Table VI image (p75) and the VI.G text (p43). In Column 4 (relative mobility) the Gini (0.006, SE 0.071) and dropout rate (0.010, 0.064) are small relative to their SEs. Social capital is 0.154 (0.060), a ratio of about 2.6, which sits awkwardly with the text calling it not significant. In the only state-FE column (Column 2, absolute mobility only), the Gini (-0.017, 0.043) and social capital (0.065, 0.050) are small relative to their SEs. No state-FE specification is reported for relative mobility. Table VI states that SEs are clustered by state but gives no significance marks or thresholds. The authors should say which significance criterion was applied to each column and which specification supports the introduction's sentence.

Second check: another model agreed

• p. 11 Section III.A

  (which excludes 1.2% of children)

• p. 76 Online Appendix Table I, cols 3-4, 1980-1982

  85.6%

• p. 11 III.A

  strictly positive (which excludes 1.2% of children)

• p. 76 Appendix Table I cols 3-4

  93.8% 93.2%

Only the authors can say what is right here.

Judge's note: Appendix Table I (p76) columns 3 and 4 are percentages of cohort size. Weighting the 1980-82 rows by cohort size, the drop is about 54.7k out of about 9.87m matched children, roughly 0.55%. Allowing for one-decimal rounding of each row still gives well under 1%. The 1980-91 total row gives 1.2/96.0, about 1.25%. So the figure most likely refers to the extended sample; the authors should confirm.

Second check: another model agreed

• p. 19 Section IV.B, Lifecycle Bias

  varying the 5-year window used to measure parent income from a starting year of 1996

• p. 94 Online Appendix Figure II notes, Panel B

  The last point measures income in 2010, when parents are 55.

Only the authors can say what is right here.

Judge's note: Seen on p19 and p94. A 5-year window starting in 2010 would need data through 2014, beyond the 2012 end of the data. The figure note reads as single-year measurement. The authors should state what window each point uses.

Second check: another model agreed

• p. 20 Section IV.B

  on parent income rank in each year from 1996 to 2012

• p. 94 Online Appendix Figure II notes, Panel C

  varying the year in which parent income rank is measured from 1996 to 2011

• p. 20 Section IV.B, Lifecycle Bias

  in each year from 1996 to 2012

• p. 20 Section IV.B

  virtually constant when children are between ages 3 and

Judge's note: Panel C on p94 plots points from age 3 to 18, i.e. 1996-2011 for the 1993 cohort. The text on p20 and footnote 28 say 1996-2012 and ages 3-19. The two descriptions of the same series differ by one year.

Second check: another model agreed

• p. 21 Alternative Income Definitions

  falls by approximately 10%, from 0.341 to 0.312

• p. 70 Table I row 6 col 1

  0.312

• p. 21

  The rank-rank correlation falls by approximately 10%, from 0.341 to 0.312

Judge's note: Text on p.21 and Table I rows 4 and 6, col 1 (0.341, 0.312) agree. The relative fall works out to 8.5%, so 'approximately 10%' is loose wording. The numbers themselves are consistent.

Second check: another model agreed

• p. 28 Section V.C

  Cleveland – approxi- mately 100 miles away – is ranked in the bottom 10

• p. 72 Table III

  40 Cleveland, Ohio

• p. 72 Table III row 40

  Cleveland, Ohio

Judge's note: Table III (p.72) lists Cleveland at rank 40 (38.2), tied in displayed value with Nashville (41) and New Orleans (42). Because of the tie it is borderline, but the table's own ranking puts it 11th from the bottom.

Second check: another model agreed

• p. 38 Section VI.B

  the correlation between commute times and upward mobility is 0.605

Judge's note: p.38: the preceding sentences define the variable correctly and state that shorter commutes go with higher mobility, so the context resolves it. The shorthand 'commute times' still reverses the sign if read literally.

Second check: another model agreed

• p. 41 Section VI.D

  very highly correlated with upward mobil- ity across all specifications

• p. 83 Online Appendix Table VIII, Test Score row, col 3

  0.176

Only the authors can say what is right here.

Judge's note: Figure VIII (p.91) shows only unweighted correlations. Appendix Table VIII (p.83) gives a test-score correlation of 0.176 (SE 0.220) in the pop-weighted column, while dropout is -0.433. 'Across all specifications' overstates the result for test scores.

Second check: another model agreed

• p. 53 Appendix D

  the implied between-group IGE is 78.8/79.9 = 0.986

Only the authors can say what is right here.

Judge's note: I reworked the step from p.53. With a black indicator as instrument, the IV slope is the child-generation gap divided by the parent-generation gap. In levels that is 20.1/21.2 = 0.948. In logs it is ln(0.799)/ln(0.788), about 0.942. The quotient 78.8/79.9 = 0.986 is computed correctly but does not follow from the stated definition. The qualitative point (high persistence, consistent with Clark) stands either way, so the authors should say which computation they intend.

Second check: another model agreed

• p. 59 Appendix G, Taxes

  difference between 2008 state income tax rates for incomes above $100,000 and incomes in the bottom tax bracket

• p. 61 Appendix H

  difference between the top state income tax rate and the state income tax rate for individuals with taxable income of $20,000 in 2008

• p. 59

  rates for incomes above $100,000 and incomes in the bottom tax bracket

Only the authors can say what is right here.

Judge's note: Both passages are confirmed on the page images (p.59 and p.61). The reference points differ ($100,000 vs the top rate; bottom bracket vs $20,000). The two may coincide in practice, but only the authors can say which definition was used.

Second check: another model agreed

• p. 59 Appendix G

  calculate mean EITC rate for the years 1980-2001

• p. 61 Appendix H

  mean state EITC rate between 1981 and 2001

• p. 59 Appendix G, Taxes

  mean EITC rate for the years 1980-2001

Only the authors can say what is right here.

Judge's note: Both passages are confirmed on the page images (p.59 and p.61). The difference is harmless if no state had an EITC in 1980, but the paper gives no reason for it, so the authors should say which window was used.

Second check: another model agreed

• p. 76 Online Appendix Table I, col 4 rows 1980-1982

  85.2%

• p. 70 Table I, Number of observations col 1

  9,867,736

Only the authors can say what is right here.

Judge's note: Recomputed from App. Table I (p76): Column 2 reproduces the full-sample count 11,262,459 of App. Table IV, which validates the method. Column 4 (85.2%, 91.1%, 93.2%) gives about 9,814k and Column 3 gives about 9,869k, against 9,867,736 reported. Rounding of the percentages and cohort sizes explains only a few thousand, not about 54k. The text's statement that the positive-income restriction 'excludes 1.2% of children' (p11) also exceeds the roughly 0.55% drop from Column 3 to Column 4. No differing definition is given anywhere (parent age 15-40 is part of both). Only the authors can reconcile these.

Second check: another model agreed

• p. 78 Online Appendix Table III, Children

  Fraction Single

• p. 79 Online Appendix Table IV, col 2

  44.31%

Only the authors can say what is right here.

Judge's note: Both values confirmed on images p78 and p79. Under the stated definitions the two categories are near-complements, yet they sum to only 88.6%. The SD of 49.7% fits p=0.443 or p=0.557 equally, so it cannot tell which label holds. A switch rate of about 11% between categories is possible if 'married' requires joint filing in both years, but the exact coincidence suggests one label may be inverted. Only the authors can say.

Second check: another model agreed

• p. 79 Online Appendix Table IV, Observations cols 4 and 8

  194,501

Only the authors can say what is right here.

Judge's note: Image p79 confirms the values. The notes say the weight sum is the average of the 2011 and 2012 weight sums and mention no rescaling, so samples differing by 3,940 observations would be expected to have different weight sums. The CPS columns have identical N and weights. One of the figures may be inconsistent; only the authors can say which.

Second check: another model agreed

• p. 8 Section II

  (Dahl and Deleire 2008)

• p. 3 Introduction

  Dahl and DeLeire (2008)

Judge's note: The p8 image shows "Deleire".

Second check: none (judge only)

• p. 4 Introduction, first line

  correlation between children’s and parents positions in the income distribution

Judge's note: The p4 page image shows no apostrophe after "parents".

Second check: none (judge only)

• p. 5 Introduction, paragraph on college attendance and teenage birth gradients

  rather than after they enter labor market

Judge's note: Confirmed on the p5 image.

Second check: none (judge only)

• p. 9 Section II, Absolute Mobility

  children whose grow up in low-income

• p. 31 V.D

  yields very spatial patterns

• p. 39 VI.C

  very highly strongly correlated

• p. 11 III.A

  begin to the leave the household

Judge's note: All three are visible on the page images (p9, p31, p39). The "to the leave" slip is judged separately under C010.

Second check: none (judge only)

• p. 11 Section III.A

  many children begin to the leave

• p. 11 Section III.A

  many children begin to the leave the household

• p. 11

  many children begin to the leave the household starting at age 17

Judge's note: Confirmed on the p11 image.

Second check: none (judge only)

• p. 14 Footnote 18

  55.1/2+0.02=27.7%

• p. 14

  We assign these children a rank of 55.1/2+0.02=27.7%

Judge's note: Seen on the p14 image. The printed result and the text both fit +0.2, so "0.02" is a typo.

Second check: none (judge only)

• p. 31 Section V.D

  yields very spatial patterns

• p. 31

  wage earnings instead of total income yields very spatial patterns

Judge's note: The p.31 image shows the phrase as quoted.

Second check: none (judge only)

• p. 39 Section VI.C

  remains very highly strongly correlated

• p. 39 Section VI.C, Table V discussion

  remains very highly strongly correlated with upward mobility

Judge's note: The p.39 image shows the phrase as quoted.

Second check: none (judge only)

• p. 47 Appendix A, SOI sample

  523,700 total observations

• p. 77 Online Appendix Table II, Total

  523,699

Judge's note: Summed column 1 of Appendix Table II (p.77) myself: 523,699. The unique-children total 228,295 matches in both places.

Second check: none (judge only)

• p. 51 Appendix C

  it is not sensitive small changes

• p. 52 Appendix C end

  which of these differences in responsible

• p. 52 Appendix C end

  which of these differences in responsible for the differences

Judge's note: I checked both phrases on the page images for p.51 and p.52. Both are printed exactly as quoted.

Second check: none (judge only)

• p. 51 Appendix C Empirical Estimates

  Between the 10th and 90 percentiles

Judge's note: The crop of p.51 shows '10th and 90 percentiles' as printed.

Second check: none (judge only)

• p. 52 Appendix D

  names held by fewer that 25, 50, or 100 children

• p. 99 Online Appendix Figure VII notes

  Areas with fewer that 250 children

Judge's note: The typo is visible on the p.52 crop and in the Figure VII notes on p.99. The same wording also appears in the Figure VIII notes on p.100.

Second check: none (judge only)

• p. 54 Appendix E

  We simply follows this recommendation.

Judge's note: Visible on the p.54 crop.

Second check: none (judge only)

• p. 59 Appendix G, Family Structure

  These data from from the

• p. 59 Appendix G, Family Structure

  These data from from the 2000 census

• p. 59

  These data from from the 2000 census long

• p. 59

  children present. These data from from the 2000 census lon

Judge's note: Visible on the p.59 crop.

Second check: code check

• p. 67 References, Hotz and Scholz

  Rober A. Moffitt

Only the authors can say what is right here.

Judge's note: Rests on outside knowledge: the editor of Means-Tested Transfer Programs in the United States is Robert A. Moffitt. The p.67 crop shows 'Rober A. Moffitt' as printed.

Second check: none (judge only)

• p. 68 References, Mitnik et al. 2014

  Intergnerational Income Mobility

• p. 67 References, Hotz and Scholz

  in Rober A. Moffitt

• p. 68

  Estimates of Intergnerational Income Mobility Using Administrative Data.

Judge's note: Page images p68 and p67 show both misspellings exactly as quoted: 'Intergnerational' for 'Intergenerational' and 'Rober' for 'Robert'.

Second check: none (judge only)

• p. 72 Table III notes

  (see Appendix F for details). . See Online Data Table V

Judge's note: The stray '. .' is visible on the page image p72, so it is not an extraction artefact.

Second check: none (judge only)

• p. 96 Online Appendix Figure IV panel title

  Minneapolis, IN

Judge's note: The crop of p96 shows 'Minneapolis, IN'. The paper's own Table III lists the CZ as 'Minneapolis, Minnesota', so the state abbreviation is wrong.

Second check: none (judge only)

• p. 97 Online Appendix Figure V notes

  Areas with fewer that 250 children

• p. 98 Online Appendix Figure VI notes

  Areas with fewer that 250 children

• p. 99 Online Appendix Figure VII notes

  Areas with fewer that 250 children

• p. 100 Online Appendix Figure VIII notes

  fewer that 250 children

Judge's note: The page images show the typo in the Figure V notes (p.97) and the Figure IX notes (p.101). The text layer also has it at p.98, p.99 and p.100 (text lines 5765, 5782, 5801), and once more in the Appendix text at line 2046 ("names held by fewer that 25, 50, or 100 children"). This is a cosmetic misspelling.

Second check: none (judge only)