[1] Source: see Data and Methodology
[2] Gladwell cites a study showing that the oldest 4th graders score between 4 and 12 percentiles than the youngest.
[3] The study referred to in note 2 found that among undergraduates at American four-year colleges, 88% as many students are born in December as in January.
[4] Source: Baseball Almanac
[5] Source: “CEO Compensation” (2010), Frydman and Jenter
[6] Source: see Data and Methodology
[7] Source: Payscale website
[8] There is evidence, however, that we can improve performance on IQ tests by working out regularly at a mental gym. For example, regular practice at “n-back” memorization drills has been shown to boost IQ scores by statistically significant amounts.
[9] Source: recent research at Harvard’s Kennedy School of Government
[10] Source: Duncan, Greg and Murnane
[11] Source: see Data and Methodology
[12] Source: “The Illusions of Entrepreneurship”, p. 99
[13] Source: Kaplan and Rauh
[14] Source: see Data and Methodology
[15] This calculation is explained in Data and Methodology
[16] The condition for the income elasticity and the correlation coefficient to be the same is that the standard deviation of the distribution of parents’ income must be the same as the standard deviation of the distribution of children’s income. In an age of increasing income inequality, this is probably not true—the standard deviation for children is probably getting wider than that for parents. On the other hand, if increasing inequality is accompanied by increasing “rigidity,” that is, if the power of the transmission vectors is increasing, this will tend to increase the parent-child correlation over time. Lastly, as mentioned, the correlation would be higher if earnings were measured for more than 16 years
[17] Source: see Data and Methodology
[18] Source: see Data and Methodology
[19] If you don’t believe that PBS is an important part of the social safety net, then read this op-ed by Charles Blow.
[20] Source: see Data and Methodology
[21] Source: see Data and Methodology
[22] These calculations assume that the rates apply to 2009 Adjusted Gross Income with no tax credits. This is of course a counter-factual assumption.
Data and Methodology
Calendar Lottery
Figure 1: The source data for Figure 1 are found in the table “PINC-03. Educational Attainment--People 25 Years Old and Over, by Total Money Earnings in 2011, Work Experience in 2011, Age, Race, Hispanic Origin, and Sex” in U.S. Census Bureau, Current Population Survey, 2012 Annual Social and Economic Supplement. The sample is 163,000 people, with 25% falling into each of four age groups: 25-34, 35-44, 44-54, and 55 and over. Since no income amounts are reported for the 7.3% of the sample with incomes above $100,000, the average 2009 Adjusted Gross Income for individuals with AGI above $100,000 was used ($205,000).
Gender Lottery
Figure 4.1 and Figure 4.2: These figures are based on the study “Education and Synthetic Work-Life Earnings Estimates” published by the US Census Bureau. Source data for this study are from the Census Bureau's Multiyear American Community Survey (ACS). “ESWLE” used data for 2006, 2007 and 2008. The Census Bureau collects the ACS data each year by interviewing approximately 2 million households.
The ESWLE methodology sorts individuals between the ages of 25 to 65 into eight age groups of five years each, nine education level groups, five ethnic groups, and two gender groups. The average “synthetic work-life earnings” for each education/ethnic/gender group (9x5x2=90 groups in all) is then calculated as the sum of the earnings for each of the eight age classes in that group. For example, the lifetime earnings for females with a K-12 education are estimated by summing the average annual earnings for K-12 black females between age 25 and 29, age 30 and 34, age 35 and 39, and so forth. This sum is then multiplied by five, since there are five years in each age group. The result is the estimated earnings over forty years for a K-12 female
Table 2-C in ESWLE (“Median Synthetic Work-Life Earnings by Education, Race/Ethnicity, and Gender: All Persons”) shows the estimated median 40-year earnings for each of the 45 groups of males and the 45 groups of females. For Figure 4.1, these groups were ranked from lowest to highest median group earnings. Using this distribution, the weighted average of the median earnings for each 10% of males and females is calculated. These values are plotted in Figure 1. It should be noted that this method does not yield a representation of the full distribution of earnings, because it is based on the median values for each of the 90 groups reported in ESWLE, and so obscures the underlying variation within each of those 90 groups.
The source data for Figure 4.2 are found in Table 2-A in “Education and Synthetic Work-Life Earnings Estimates,” which shows median earnings for full-time workers rather than for all people as in Table 2.C. The decile values for males and females in Figure 4.2 are developed in the same way as those in Figure 4.1.
Education Lottery
Figure 10: This is derived from the interquartile ranges of for lifetime earnings presented in Table 1A of “The College Payoff: Education, Occupations, Lifetime Earnings.” The authors estimated lifetime earning using the same source data and a similar methodology to that described for Figures 4.1. and 4.2. Since earnings are assumed to be log-normally distributed, the natural logs of the interquartile ranges were used to calculate standard deviations for each education class. These were used to calculate the distributions reproduced as Figure 10.
Why We Have What We Have
Figure 12: The source data for this figure are the same as described above for Figure 1. From these data, the mean and standard deviation for each gender was computed, and these were used to compute the values in Figure 12, assuming that earnings are log-normally distributed.
The correlation of 0.22 between gender and earnings was calculated from the same source data as Figure 12, except that those reporting no income were included in this calculation. When one variable can take on only one of two values (gender in this case), and the other can take on a range of values (earnings), the appropriate statistical algorithm for calculating the correlation between the two variables is the “point biserial” correlation coefficient. The algorithm for this is the difference between the mean for males and the mean for females, divided by the population standard deviation, multiplied by the square root of the product of the male fraction times the female fraction.
The values for mean male earnings, mean female earnings and standard deviation for the full sample can be calculated from the source data described for Figure 1.
Figure 13: The source data for this figure are taken from Table 5.10 in Hertz (2005). These data were used to calculate the probability that a child born to parents in the lower five deciles (“poor children”) will have lifetime earnings in each of the ten children's earnings deciles. The same calculation is performed for the child born to parents in the upper five deciles (“rich children”). These probability distributions are then transformed into earnings distributions shown in Figure 13 using the earnings values at each decile reported in the source data described above for Figure 12.
Figure 14: Figure 14 combines the source data and calculations described for Figure 12 and Figure 13. The median for “rich boys” in Figure 14 was calculated as the median value for “rich children” from Figure 13 times the ratio of the median value for men (Figure 12) to median value for the entire population (calculated as described above in the discussion of Figure 12). The standard deviation for “rich boys” was calculated as the median for rich boys times the average of the ratios of the standard deviation to the mean for rich children (calculated as described in the discussion of Figure 13) and for men (calculated as described in the discussion of Figure 12.)
Calculations of probabilities that a “poor girl's” or “rich boy's” earnings will exceed a specified value are p
erformed using the natural logs of the earnings distributions described for Figure 13. For example, the log of the median annual income for “poor girls” is 9.24, and the log of the “poor girls” standard deviation of 1.02. The log of the median income for “rich boys” is 10.69. The difference between the log of the “rich boy” and “poor girl” medians is 10.69-9.24 = 1.45, which is 1.42 “poor girl” standard deviations. From the probability density function of the normal distribution, we find that 1.42 standard deviations corresponds to a probability density of 92.2%, which means there is a 7.8% probability that a “poor girl's” earnings will exceed those of a “rich boy.”
How Government Should Take What it Needs for Charity
Figure 15: The source data for Figure 15 are the same as described for Figure 1. Figure 15 plots the values for each income bracket from lowest to highest (Y axis) against the percentage of the sample falling within each income bracket (X axis).
Figure 18: The source data for Figure 18 is Adjusted Gross Income for 2009 as reported in the IRS publication “Statistics of Income Division, July 2011, Table 1.1, Selected Income and Tax Items, by Size and Accumulated Size of Adjusted Gross Income, Tax Year 2009.” The line “current rates” in Figure 18 is derived by applying the marginal tax rates for 2012 to this 2009 AGI data to calculate the average tax rate for each AGI bracket. “Fortune Tax rates” are the average rates that result when all 2009 AGI is taxed at the lesser of 10% of AGI or 50% of the amount by which AGI exceeds the 65th percentile amount of $38,750. “Age of Affluence” average rates are calculated using the marginal tax rates for 1970, with tax brackets adjusted for inflation.
Sources
Bedard and Dhuey, 2006. “The Persistence of Early Childhood Maturity: International Evidence of Long-Run Age Effects,” Quarterly Journal of Economics 121 no. 4
Carnevale, Anthony P., Stephen J. Rose, and Ban Cheah, “The College Payoff: Education, Occupations, Lifetime Earnings.” Georgetown University Center on Education and the Workforce, https://www9.georgetown.edu/grad/gppi/hpi/cew/pdfs/collegepayoff-complete.pdf
Duncan, Greg and Murnane, 2011. Whither Opportunity? Rising Inequality, Schools, and Children's Life Chances.
Freeland, Chrystia: “Super-Rich Irony.” The New Yorker, October 8, 2012.
Gladwell, Malcolm: “Outliers: The Story of Success.” Back Bay Books, Little Brown, 2008
Hertz, Tom, 2005. “Rags, Riches and Race: The Intergenerational Economic Mobility of Black and White Families in the United States,” in Unequal Chances: Family Background and Economic Success, edited by Bowles, Gintis and Groves, Russell Sage Foundation, New York, Princeton University Press
Julian, Tiffany A. and Robert A. Kominski, 2011. “Education and Synthetic Work-Life Earnings Estimates.” American Community Survey Reports, ACS-14. U.S. Census Bureau, Washington, DC. https://www.census.gov/prod/2011pubs/acs-14.pdf
Kaplan and Rauh, 2007. “Wall Street and Main Street: What Contributes to the Rise in the Highest Incomes?” National Bureau of Economic Research
Levitin, Daniel J., This is Your Brain on Music: The Science of Human Obsession (New York: Dutton, 2006), p.197
Lewis, Michael: “Don't Eat Fortune's Cookie,” Princeton University's 2012 Baccalaureate Remarks.
Mazumder, Bhashkar, 2005.“The Apple Falls Even Closer to the Tree than We Thought: New and Revised Estimates of the Intergenerational Inheritance of Earnings” in Unequal Chances: Family Background and Economic Success, edited by Bowles, Gintis and Groves, Russell Sage Foundation, New York, Princeton University Press
Murray, Charles, 1998. “Income Inequality and IQ.” The AEI Press, Washington DC
Putnam, Robert D., 2012. “Requiem for the American Dream? Unequal Opportunity in America.” Slides presented at Aspen Ideas Festival. The Saguaro Seminar, Kennedy School of Government, Harvard University.
Schmidt, Frank L.,2009. “Select on Intelligence,” in Handbook of Principles of Organizational Behavior¸edited by Edwin A. Locke, 2nd edition. John Wiley and Sons.
Slaughter, Anne-Marie. “Why Women Still Can't Have It All.” The Atlantic, July/August, 2012.
Shane, L. Scott, 2008. The Illusions of Entrepreneurship: the Costly Myths that Entrepreneurs, Investors and Policymakers Live By. Yale University Press
About the Author
Daniel Badger drew top tickets in the birth lotteries in 1946. He did even better in the education lottery, attending Greenwich Country Day School, Phillips Academy Andover, Yale University, Cambridge University, and Harvard's Kennedy School of Government--a total of fifteen years of private education. His tickets in the job lottery were also pretty good, although two employers ran aground. He worked at the U.S. Regulatory Commission, the U.S. Department of Energy, and the International Energy Agency, a couple of energy consulting firms, Enron, and Babcock & Brown. He currently works in London for a financial advisory firm specializing in renewable energy.
Readers' comments are welcome, and can be emailed to Mr. Badger at [email protected].
The Grasshopper and the Ant, or the Beautiful and the Damned? Why We Have What We Have, and How Government Should Take What it Needs for Charity Page 6