_notes.txt

<strong>Table 1.1. Key labor market indicators and living-standards benchmarks, 2000–2011. </strong>Underlying data are from the Current Population Survey (CPS) public data series; the CPS Annual Social and Economic Supplement microdata and <em>Historical Income Tables</em> Table H-5, “Race and Hispanic Origin of Householder–Households by Median and Mean Income: 1967 to 2010”; CPS Outgoing Rotation Group microdata (see Appendix B for details on CPS-ORG microdata); the Bureau of Labor Statistics Current Employment Statistics; and unpublished Total Economy Productivity data from the Bureau of Labor Statistics Labor Productivity and Costs program.

<strong>Table 1.2. Key labor market indicators and living-standards benchmarks, 1979–2011. </strong>See note for Table 1.1.

<strong>Table 1.3. Middle-fifth household income, minus selected key sources, 1979–2007. </strong>Underlying data for income, transfers, and pensions are from the Congressional Budget Office Web resource, <em>Average Federal Taxes by Income Group</em>, “Sources of Income for All Households, by Household Income Category, 1979 to 2007” [Excel spreadsheet] and unpublished data related to the resource. Underlying data for health care deflation are from the Bureau of Labor Statistics <em>Consumer Price Indexes </em>database. Underlying data for hours worked are from Current Population Survey Annual Social and Economic Supplement microdata; see Appendix A for details on microdata. Income data are deflated using a health care deflator, and then the contributions of additional transfers, hours worked, and pensions since 1979 are taken out in sequence. Note that the unpublished CBO data are unrounded, and produce slightly different income dollar values and thus an income growth rate for the middle fifth (19.1 percent) that differs by .1 percentage point from the income growth rate from the rounded, publicly available CBO data underlying Figure 1I. Note that the “hours worked” increases in some periods because total earnings in the CBO data dropped <em>more </em>than hourly earnings in the CPS data (which is where the hourly earnings are measured from) over this period. This implies that hours dropped more than hourly earnings over this period in the CBO data. In other words, if you remove the effect of hours (i.e., leave only the effect of hourly earnings), total earnings will rise.

<strong>Table 1.4. Employer-provided health insurance and pension coverage, by race and ethnicity, 1979–2010. </strong>Underlying data are from Current Population Survey Annual Social and Economic Supplement microdata; see Appendix A for details.

<strong>Table 1.5. Employer-provided health insurance and pension coverage, by gender, 1979–2010. </strong>Underlying data are from Current Population Survey Annual Social and Economic Supplement microdata; see Appendix A for details.

<strong>Figure 1A. Payroll employment and the number of jobs needed to keep up with the growth in the potential labor force, Jan. 2000–Dec. 2011. </strong>Underlying data are from the Bureau of Labor Statistics Current Employment Statistics public data series and a 2012 Congressional Budget Office report, <em>The Budget and Economic Outlook</em>, Table 2-3, “Key Assumptions in the CBO’s Projection of Potential GDP.” Since the CBO estimates of the size of the potential labor force are annual, the annual values are assigned to June of each year and extrapolated for the monthly figure.

<strong>Figure 1B. Home prices and their impact on residential investment and housing wealth, 1995–2011. </strong>Underlying data are from Shiller (2005 and 2012), Bureau of Economic Analysis National Income and Product Accounts,Table 1.1.5, “Gross Domestic Product,” and Federal Reserve Board (2012), Flow of Funds Accounts of the United States. Home prices are indexed such that 1997=100, and residential investment and the wealth effect on consumption are relative to 1997 average as a share of GDP.

<strong>Figure 1C. Employment-to-population ratio, age 25–54, Jan. 1995–Dec. 2011. </strong>Underlying data are from the Current Population Survey public data series.

<strong>Figure 1D. Unemployment rate and real median wage decline, 1991–2011. </strong>Underlying data for the unemployment rate are from the Current Population Survey public data series. The unemployment rate is lagged by one year in the figure. Underlying data for median wages are from CPS Outgoing Rotation Group microdata; see Appendix A for details.

<strong>Figure 1E. Change in real family income of the middle fifth, actual and predicted, 2000–2018. </strong>Underlying data are from the Current Population Survey public data series on unemployment and from CPS Annual Social and Economic Supplement <em>Historical Income Tables</em>, Table F-2, “Share of Aggregate Income Received by Each Fifth and Top 5 Percent of All Families, All Races: 1947– 2010”; Table F-3, “Mean Income Received by Each Fifth and Top 5 Percent of Families, All Races: 1966 to 2010”; and Table F-5, “Race and Hispanic Origin of Householder—Families by Median and Mean Income.” Real family income is indexed such that 2000=100. The projections are based on a regression analysis, based roughly on Katz and Krueger (1999), that uses the annual change in inflation-adjusted income of families in the middle fifth of the money income distribution as the dependent variable and the level of unemployment as the independent variable. The projections then use the regression parameters to forecast annual changes in middle-fifth family income based on unemployment forecasts through 2018 that are made by the Congressional Budget Office and Moody’s Economy.com, a division of Moody’s Analytics.

<strong>Figure 1F. Cumulative change in total economy productivity and real hourly compensation of selected groups of workers, 1995–2011.</strong> Productivity data, which measure output per hour of the total economy, including private and public sectors, are from an unpublished series available from the Bureau of Labor Statistics Labor Productivity and Costs program on request. Wage measures are the annual data used to construct tables in Chapter 4: median hourly wages (at the 50th percentile) from Table 4.4 and hourly wages by education from Table 4.14. These are converted to hourly compensation by scaling by the real compensation/wage ratio from the Bureau of Economic Analysis National Income and Product Accounts (NIPA) data used in Table 4.2.

<strong>Figure 1G. Share of total household income growth attributable to various income groups, 1979–2007. </strong>Underlying data are from the Congressional Budget Office <em>Average Federal Taxes by Income Group</em>, “Sources of Income for All Households, by Household Income Category, 1979 to 2007” [Excel spreadsheet]. Each group’s contribution to overall income growth is calculated by multiplying the change in its average income from 1979 to 2007 by its share of the distribution (where, for example, the share of the distribution for the top 1 percent is .01), and dividing the result by the change in overall average income growth over the same time period. For pretax income calculations of the 90th–&lt;95th percentile and 95th–99th percentile, see Figure 2M notes.

<strong>Figure 1H. Share of average income growth accounted for by the top 5 percent and top 1 percent, by dataset and income concept, 1979–2007. </strong>Underlying data are from Piketty and Saez (2012, Table A-6); Congressional Budget Office, <em>Average Federal Taxes by Income Group</em>, “Sources of Income for All Households, by Household Income Category, 1979 to 2007” [Excel spreadsheet]; and Burkhauser, Larrimore, and Simon (2011), Table 4, “Quintile Income Growth by Business Cycle Using Each Income Series.” Each income concept’s contribution to overall income growth is calculated by multiplying the change in its average income from 1979 to 2007 by its share of the distribution (where, for example, the share of the distribution for the top 1 percent is .01,) and dividing the result by the change in overall average income growth over the same time period.

<strong>Figure 1I. Change in real annual household income, by income group, 1979–2007. </strong>Underlying data are from the Congressional Budget Office, <em>Average Federal Taxes by Income Group</em>, “Sources of Income for All Households, by Household Income Category, 1979 to 2007” [Excel spreadsheet]. Cumulative growth is calculated by dividing the average pretax income in the base year (1979) into average pretax income in each subsequent year (1980–2007). The data provide average pretax income for the bottom, second, middle, fourth, and top fifths, and for the top 10, 5, and 1 percent. For the 80th–&lt;90th percentile, average pretax income is calculated by subtracting the aggregate income of the top 10 percent from aggregate income of the top fifth and dividing by the total number of households in the 80th–&lt;90th percentile. Aggregate income is calculated by multiplying the number of households in each income group by average pretax income. The number of households is calculated by subtracting the number of households in the top 10 percent from the number of households in the top fifth. This same procedure is done between the top 10 percent and top 5 percent to calculate average pretax income for the 90th–&lt;95th percentile and between the top 5 percent and top 1 percent to calculate the average pretax income for the 95th–&lt;99th percentile. Data are inflated to 2011 dollars using the CPI-U-RS and then indexed to 1979=0. Note that this publicly available CBO dataset is rounded, and produces slightly different income dollar values and thus an income growth rate for the middle fifth (19.2 percent) that differs by .1 percentage point from the income growth rate from the unpublished, unrounded CBO data underlying Table 1.3.

<strong>Figure 1J. Average family income growth, by income group, 1947–2007.</strong> CPS Annual Social and Economic Supplement <em>Historical Income Tables</em>, Table F-2, “Share of Aggregate Income Received by Each Fifth and Top 5 Percent of All Families, 1947–2010”; Table F-3, “Mean Income Received by Each Fifth and Top 5 Percent of Families, All Races: 1966–2010”; and Table F-5, “Race and Hispanic Origin of Householder—Families by Median and Mean Income.” Data are inflated to 2011 dollars using the CPI-U-RS.

<strong>Figure 1K. Income of middle-fifth households, actual and projected assuming growth equal to growth rate of overall average income, 1979–2007. </strong>Underlying data are from the Congressional Budget Office <em>Average Federal Taxes by Income Group</em>, “Sources of Income for All Households, by Household Income Category, 1979 to 2007” [Excel spreadsheet]. Data for the middle fifth are shown as is and when applying the cumulative growth rate of the average income for all households.

<strong>Figure 1L. Cumulative change in real annual wages, by wage group, 1979–2010. </strong>Data taken from Kopczuk, Saez, and Song (2010), Table A-3. Data for 2006 through 2010 are extrapolated from 2004 data using changes in wage shares computed from Social Security Administration wage statistics (data for 2010 are at http://www.ssa.gov/cgi-bin/netcomp.cgi). The final results of the paper by Kopczuk, Saez, and Song printed in a journal used a more restrictive definition of wages so we employ the original definition, as recommended in private correspondence with Kopczuk. SSA provides data on share of total wages and employment in annual wage brackets such as for those earning between $95,000.00 and $99,999.99. We employ the midpoint of the bracket to compute total wage income in each bracket and sum all brackets. Our estimate of total wage income using this method replicates the total wage income presented by SSA with a difference of less than 0.1 percent. We used interpolation to derive cutoffs building from the bottom up to obtain the 0–90th percentile bracket and then estimate the remaining categories. This allows us to estimate the wage shares for upper wage groups. We use these wage shares computed for 2004 and later years to extend the Kopczuk, Saez, and Song series by adding the changes in share between 2004 and the relevant year to their series. To obtain absolute wage trends we used the SSA data on the total wage pool and employment and computed the real wage per worker (based on their share of wages and employment) in the different groups in 2011 dollars.

<strong>Figure 1M. Intergenerational correlations between the earnings of fathers and sons in OECD countries.</strong> The figure is adapted from Corak (2011), Figure 1, “Comparable Estimates of the Intergenerational Elasticity between Father and Son Earnings for the United States and Twenty Four Other Countries.” “Earnings” refers to wages.

<strong>Figure 1N. Elasticities between parental income and sons’ earnings, 1950–2000.</strong> Data are from Aaronson and Mazumder (2007), Table 1,“Estimates of the IGE Using Census IPUMS Data.” Data reflect annual family income for the parents and annual earnings for the sons.

<strong>Figure 1O. Unemployment rate, by race and ethnicity, 1979–2011.</strong> Underlying data are basic monthly Current Population Survey microdata. As with other CPS microdata analyses presented in the book, race/ethnicity categories are mutually exclusive (i.e., white non-Hispanic, black non-Hispanic, and Hispanic any race).

<strong>Figure 1P Median wealth by race, 1983–2010. </strong>Underlying data are from the 2010 Survey of Consumer Finances (SCF) data prepared in 2012 by Edward Wolff for the Economic Policy Institute. The definition of wealth used in this analysis of the SCF is the same definition of wealth used in the analysis of the SCF conducted by Bricker et al. (2012), except that the Bricker et al. analysis includes vehicle wealth, while this analysis does not.

<strong>Table 2.1. Average family income, by income group, 1947–2010.</strong> Underlying data are from Current Population Survey Annual Social and Economic Supplement <em>Historical Income Tables</em>, Table F-2, “Share of Aggregate Income Received by Each Fifth and Top 5 Percent of All Families, All Races: 1947– 2010,” Table F-3, “Mean Income Received by Each Fifth and Top 5 Percent of Families, All Races: 1947 to 2010,” and Table F-5, “Race and Hispanic Origin of Householder—Families by Median and Mean Income: 1947 to 2010.” The years 1947, 1979, 1989, 2000, and 2007 are highlighted throughout the chapter because they are employment cycle peaks and are similar in nature to business cycle peaks. 1995 represents a midway point between cycles to show the growth or stagnation of the period. 2010 is highlighted because it is the most recent year for which data are available. Data are inflated to 2011 dollars using the CPI-U-RS (Consumer Price Index Research Series Using Current Methods).

<strong>Table 2.2. Average household income, by income group, 1967–2010.</strong> Underlying data are from Current Population Survey Annual Social and Economic Supplement <em>Historical Income Tables</em>, Table H-3, “Mean Income Received by Each Fifth and Top 5 Percent, All Races: 1967 to 2010.” Data are inflated to 2011 dollars using the CPI-U-RS.

<strong>Table 2.3. Minimum income thresholds for family and household income, by income group, 1947–2010.</strong> Underlying data are from Current Population Survey Annual Social and Economic Supplement <em>Historical Income Tables</em>, Table F-1, “Income Limits for Each Fifth and Top 5 Percent of Families (All Races): 1947 to 2010,” and Table H-1, “Income Limits for Each Fifth and Top 5 Percent of All Households: 1967 to 2010.” Data are inflated to 2011 dollars using the CPI-U-RS.

<strong>Table 2.4. Sources of pretax comprehensive income, by income group, 2007.</strong> Underlying data are from the Congressional Budget Office, <em>Average Federal Taxes by Income Group,</em> “Sources of Income for All Households, by Household Income Category, 1979 to 2007” [Excel spreadsheet]. Shares of pretax income, by income source, are given by CBO for the bottom, second, middle, fourth, and top fifth, and the top 10, 5, and 1 percent. Average pretax income is defined as the sum of each income groups’ wages, proprietors’ income, other business income, interest and dividends, capital gains pensions, cash transfers, in-kind income, imputed taxes, and other income. For the purposes of this chapter, capital income is defined as the sum of capital gains, interest and dividends, and other business income categories. Sources of income for the groups are calculated by multiplying the shares of each income source by average pretax income. To calculate average pretax income by source for the 95th–&lt; 99th percentile, the aggregate incomes of the top 5 percent were subtracted from the aggregate incomes of the top 10 percent and divided by the total number of households in the 95th–&lt;99th percentile. Aggregate income is calculated by multiplying the number of households in each income group by average pretax income source. The number of households is calculated by subtracting the number of households in the top 5 percent from the number of households in the top 10 percent. The same calculation is done for the 95th–&lt;99th percentile using the top 5 percent and the top 1 percent. The share of total income categories claimed by each group is calculated by dividing the aggregate income for each income source in each income group by the total aggregate income for all households, minus negative income. Data are inflated to 2011 dollars using the CPI-U-RS.

<strong>Table 2.5. Median family income by race and ethnicity, 1947–2010.</strong> Underlying data are from Current Population Survey Annual Social and Economic Supplement <em>Historical Income Tables</em>, Table F-5, “Race and Hispanic Origin of Householder—Families by Median and Mean Income: 1947–2010.” Unlike with CPS microdata analyses presented in the book, race and ethnicity categories are not mutually exclusive (i.e., persons of Hispanic origin may be of any race, and white and black Hispanics are counted in the white and black columns as well as the Hispanic column). Data are inflated to 2011 dollars using the CPI-U-RS.

<strong>Table 2.6. Share of average income growth accounted for by the bottom 95 percent, top 5 percent, and top 1 percent, by dataset and income concept, 1979–2007</strong>

Underlying data are from Piketty and Saez (2012, Table A-6); Current Population Survey Annual Social and Economic Supplement <em>Historical Income Tables</em>, Table H-3, “Mean Household Income Received by Each Fifth and Top 5 percent;” Congressional Budget Office <em>Average Federal Taxes by Income Group,</em> “Sources of Income for All Households, by Household Income Category, 1979 to 2007” [Excel spreadsheet]; and Burkhauser, Larrimore, and Simon (2011), Table 4, “Quintile Income Growth by Business Cycle Using Each Income Series.” Each income concept’s contribution to overall income growth is calculated by multiplying the change in its average income from 1979 to 2007 by its share of the distribution (where, for example, the share of the distribution for the top 1 percent is .01), and dividing the result by the change in overall average income growth over the same time period.

<strong>Table 2.7. Effective tax rates for selected federal taxes, by income group, 1979–2007.</strong> Underlying data are from the Congressional Budget Office,<em> Average Federal Taxes by Income Group</em>, “Average Federal Tax Rates for All Households, by Comprehensive Household Income Quintile, 1979–2007” [Excel spreadsheet]. CBO defines individual income taxes as taxes attributed directly to households paying those taxes; social insurance (payroll) taxes are taxes attributed to households paying those taxes directly or paying them indirectly through their employers. Corporate income taxes are attributed to households according to a household’s share of capital income, and federal excise taxes are attributed to households according to their consumption of the taxed good or service.

<strong>Table 2.8. Tax rate, transfer rate, and tax rate net of transfers, by income group, 1979–2007.</strong> Underlying data are from the Congressional Budget Office <em>Average Federal Taxes by Income Group</em>, “Average Federal Tax Rates for All Households, by Comprehensive Household Income Quintile, 1979–2007,” “Sources of Income for All Households, by Household Income Category 1979 to 2007” [Excel spreadsheets] and unpublished data related to the same report on the composition of in-kind income, with a breakout for health spending (both government transfers and employer-sponsored insurance benefits). The tax rate is taken directly from the first Excel spreadsheet cited here, while the transfer rate is calculated as the share of cash transfers and Medicare and Medicaid spending in comprehensive income.

<strong>Table 2.9. Educational attainment, by income group, selected years, 1979–2007.</strong> Underlying data are from Current Population Survey Annual Social and Economic Supplement microdata; see Appendix A for details. The data are sorted by household income and placed into the income groupings. Then, an hours-weighted measure of the share of all hours worked by workers with the given educational attainment is constructed for each of the income groupings.

<strong>Table 2.10. Share of market-based personal income, by income type, selected years, 1959–2010.</strong> Underlying data for total capital income, rent, dividends, interest, total labor income, wages and salaries, fringe benefits, and proprietors’ income are from Bureau of Economic Analysis National Income and Product Accounts, Table 2.1, “Personal Income and Its Disposition.” Underlying data for realized capital gains come from the Internal Revenue Service, <em>SOI Tax Stat–Individual Time Series Statistical Tables</em>, Historical Table 1, “All Individual Income Tax Returns: Sources of Income and Tax Items, Tax Years 1913–2005,” and Table 1, “Individual Income Tax Returns: Selected Income and Tax Items for Specified Tax Years, 1999–2009.” Rent, dividends, interest, total labor income, wages and salaries, fringe benefits, proprietors’ income, and net capital gains are divided by the total market income (the sum of total capital income, total labor income, and proprietors’ income) for select years.

<strong>Table 2.11. Effect of the shift from labor to capital income on the top 1 percent of households, selected years, 1979–2007.</strong> Underlying data are from the Congressional Budget Office, <em>Average Federal Taxes by Income Group</em>, “Sources of Income for All Households, by Household Income Category, 1979 to 2007” [Excel spreadsheet]. The counterfactual holds the share of all income accounted for by capital income constant at its 1979 level. By implication, this means that all non-capital income sources rise over that time period (since overall income growth is assumed to remain the same). This extra non-capital income is distributed across income groupings in proportion to their actual income shares over time. Then the counterfactual income level of the top 1 percent is calculated and compared with actual trends. Data are inflated to 2011 dollars using the CPI-U-RS.

<strong>Table 2.12. Corporate sector income shares, profit rates, and capital-to-output ratio, selected years, 1959–2010.</strong> Underlying data are from the Bureau of Economic Analysis National Income and Product Accounts, Table 1.14, “Gross Value Added of Domestic Corporate Businesses in Current Dollars and Gross Value added of Nonfinancial Domestic Corporate Business in Current and Chained Dollars” and BEA Fixed Assets Accounts, Table 6.1, “Current-Cost Net Stock of Private Fixed Assets by Industry Group and Legal Form of Organization.” Total income shares are the sum of labor and capital income, specifically the sum of line items Compensation and Net Operating Surplus to get net value added in NIPA Table 1.14. Labor share is the share of compensation in net value added and capital is net operating surplus over net value added. Pretax profit rate is the net operating surplus divided by private fixed corporate assets, line item 2 from Table 6.1. Post-tax profit rate is the net operating surplus, without taxes, divided by private fixed corporate assets. The capital-to-output ratio is private fixed corporate assets divided by the constructed net value added.

<strong>Table 2.13. Change in sources of comprehensive income, middle fifth of households, selected years, 1979–2007 (2011 dollars).</strong> Underlying data are from the Congressional Budget Office,<em> Average Federal Taxes by Income Group</em>, “Sources of Income for All Households, by Household Income Category, 1979 to 2007” [Excel spreadsheet], as well as unpublished data related to the same CBO Web resource on the composition of in-kind income, with a breakout for health spending (both government transfers and employer-sponsored insurance benefits). “Imputed taxes” are taxes that are not directly paid by households to government (such as the employer’s share of the payroll tax), but which are “paid” in the form of lower wages and thus are added by the CBO to actual, observed wages to produce the measure of “pretax” income. “Other income” in the pensions category includes withdrawals from 401 (k) plans and traditional pensions and a small category of “other income” that CBO links with pension income in its reports. Note that the unpublished CBO data are unrounded, and produce slightly different income dollar values than the publicly available CBO dataset underlying Figures 2M and 2Z. For deflation of health care benefits (both transfers and employer-provided) we use the Consumer Price Index for medical care (CPI-MC) instead of the Consumer Price Index for Urban Consumers, Research Series (CPI-U-RS) that is used throughout the book.

<strong>Table 2.14. Change in sources of comprehensive income for elderly households in the middle fifth, selected years, 1979–2007.</strong> Underlying data are unpublished data on income source by family type from the Congressional Budget Office related to its 2010 Web resource, <em>Average Federal Taxes by Income Group</em>. “Imputed taxes” are taxes that are not directly paid by households to government (such as the employer’s share of the payroll tax), but which are “paid” in the form of lower wages and thus are added by the CBO to actual, observed wages to produce the measure of “pretax” income. “Other income” in the pensions category includes withdrawals from 401 (k) plans and traditional pensions, and a small category of “other income” that CBO links with pension income in its reports. The income levels for “Wages and imputed taxes” column and the “Pensions and other income” columns are calculated by the sum of the product of the shares of wages and imputed taxes multiplied by average pre-tax income for each income group and the sum of the product of the share of pensions and other income multiplied by average pretax income. The contribution to shares from income sources is calculated by multiplying the change in the types of income sources by the changes in the total income for elderly households. Data are inflated to 2011 dollars using the CPI-U-RS.

<strong>Table 2.15. Change in sources of comprehensive income for non-elderly households in the middle fifth, selected years, 1979–2007. </strong>Underlying data are unpublished data on income source by family type from the Congressional Budget Office related to its 2010 Web resource, <em>Average Federal Taxes by Income Group</em>. “Imputed taxes” are taxes that are not directly paid by households to government (such as the employer’s share of the payroll tax), but which are “paid” in the form of lower wages and thus are added by the CBO to actual, observed wages to produce the measure of “pretax” income. “Other income” in the pensions category includes withdrawals from 401 (k) plans and traditional pensions, and a small category of “other income” that CBO links with pension income in its reports. The income levels for “Wages and imputed taxes” column and the “Pensions and other income” columns are calculated by the sum of the product of the shares of wages and imputed taxes multiplied by average pretax income for each income group and the sum of the product of the share of pensions and other income multiplied by average pretax income. The contribution to shares from income sources is calculated by multiplying the change in the types of income sources by the changes in the total income for non-elderly households. Data are inflated to 2011 dollars using the CPI-U-RS. Note that the unpublished CBO data are unrounded, and produce slightly different income dollar values than the publicly available CBO dataset underlying Figures 2M and 2Z.

<strong>Table 2.16. Contributions to middle-fifth income growth, by income category and household type, selected years, 1979–2007.</strong> Underlying data are unpublished data on income source by family type from the Congressional Budget Office related to its 2010 Web resource,<em> Average Federal Taxes by Income Group</em>. Data are inflated to 2011 dollars using the CPI-U-RS.

<strong>Table 2.17. Contribution of hours versus hourly wages to annual wage growth for working-age households, by income group, selected years, 1979–2007.</strong> Underlying data are from Current Population Survey Annual Social and Economic Supplement microdata; see Appendix A for details. Households are ranked in the same way as in the Congressional Budget Office data—by household income divided by the square root of household size. Average annual wages and annual hours worked for each income group are then calculated, and a household average for hourly wages is calculated by dividing annual wages by annual hours. Data are inflated to 2011 dollars using the CPI-U-RS.

<strong>Table 2.18. Annual hours worked by married men and women age 25–54 with children, by income group, selected years, 1979–2010.</strong> Underlying data are from the Current Population Survey Annual Social and Economic Supplement microdata; see Appendix A for details.

<strong>Table 2.19. Impact of increasing education and experience on hourly wages of individuals in the middle fifth of the income distribution, selected years, 1979–2007.</strong> Underlying data are from Current Population Survey Annual Social and Economic Supplement microdata; see Appendix A for details. Households are ranked in the same way as in the Congressional Budget Office data—by household income divided by the square root of household size. Fifty age/experience “cells” are created (five educational categories by 10 potential experience categories). Average hourly earnings are calculated for each cell. To get the counterfactual wage growth that would have happened <em>without</em> education and experience upgrading, we hold the 1979 cell weights (i.e., the shares of total hours worked in each year by a given cell) constant, but allow the within-cell wage growth to occur. Data are inflated to 2011 dollars using the CPI-U-RS.

<strong>Figure 2A. Real median family income, 1947–2010.</strong> Underlying data are from Current Population Survey Annual Social and Economic Supplement <em>Historical Income Tables</em>, Table F-5, “Race and Hispanic Origin of Householder—Families by Median and Mean Income: 1947 to 2010.” Data are inflated to 2011 dollars using the CPI-U-RS.

<strong>Figure 2B. Real median income of working-age families, 1975–2010.</strong> Underlying data are from Current Population Survey Annual Social and Economic Supplement microdata; see Appendix A for details. Data are inflated to 2011 dollars using the CPI-U-RS.

<strong>Figure 2C. Average family income growth, by income group, 1947–2007.</strong> Underlying data are from Current Population Survey Annual Social and Economic Supplement <em>Historical Income Tables</em>, Table F-3, “Mean Income Received by Each Fifth and Top 5 Percent of Families, All Races: 1966 to 2010.” Data are inflated to 2011 dollars using the CPI-U-RS.

<strong>Figure 2D. Black median family income, as a share of white median family income, 1947–2010.</strong> Underlying data are from Current Population Survey Annual Social and Economic Supplement <em>Historical Income Tables</em>, Table F-5, “Race and Hispanic Origin of Householder—Families by Median and Mean Income: 1947 to 2010.”

<strong>Figure 2E. Median family income growth, by nativity, 1993–2010.</strong> Underlying data are from Current Population Survey Annual Social and Economic Supplement microdata; see Appendix A for details. Data is inflated to 2011 dollars using the CPI-U-RS and then indexed to 1993=100.

<strong>Figure 2F. Change in average family income, by income group, 2007–2010.</strong> Underlying data are from Current Population Survey Annual Social and Economic Supplement <em>Historical Income Tables</em>, Table F-3, “Mean Income Received by Each Fifth and Top 5 Percent of Families, All Races: 1966 to 2010.” Data are inflated to 2011 dollars using the CPI-U-RS.

<strong>Figure 2G. Change in real family income from the business cycle peak years 1989, 2000, and 2007.</strong> Underlying data are from Current Population Survey Annual Social and Economic Supplement <em>Historical Income Tables</em>, Table F-3, “Mean Income Received by Each Fifth and Top 5 Percent of Families, All Races: 1966 to 2010.” Data for each recession are indexed to the business cycle peak year preceding the recession=100.

<strong>Figure 2H. Average capital gains of the top 5% of the income distribution and the S&amp;P 500 composite price index, 1979–2011.</strong> Underlying data are from Piketty and Saez (2012, Tables A-6 and A-8) and Shiller (2012). The inflation-adjusted S&amp;P 500 data are taken directly from Shiller and converted into an index (1989=100). Income derived from realized capital gains is taken from Piketty and Saez (2012) and converted into an index as well. The Shiller data can be found at: http://www.econ.yale.edu/~shiller/data.htm, and the Piketty and Saez data can be found at: http://elsa.berkeley.edu/~saez/TabFig2010.xls.

<strong>Figure 2I. Change in real median household income, by race and ethnicity, 2007–2010.</strong> Underlying data are from Current Population Survey Annual Social and Economic Supplement <em>Historical Income Tables</em>, Table H-5, “Race and Hispanic Origin of Householder—Households by Median and Mean Income: 1967–2010.”

<strong>Figure 2J. Change in real family income of the middle fifth, actual and predicted, 2000–2018.</strong> Underlying data are from the Current Population Survey Annual Social and Economic Supplement <em>Historical Income Tables</em>, Tables F-2, F-3, and F-5. Data are inflated to 2011 dollars using the CPI-U-RS. The projections are based on a regression analysis, based roughly on Katz and Krueger (1999), that uses the annual change in inflation-adjusted income of families in the middle fifth of the money income distribution as the dependent variable and the level of unemployment as the independent variable. The projections then use the regression parameters to forecast annual changes in middle-fifth family income based on unemployment forecasts through 2018 that are made by the Congressional Budget Office and Moody’s Economy.com, a division of Moody’s Analytics.

<strong>Figure 2K. Income growth for families at the 20th, 50th, and 95th percentiles, 1947–2010.</strong> Underlying data are from Current Population Survey Annual Social and Economic Supplement <em>Historical Income Tables</em>, Table F-1, “Income Limits for Each Fifth and Top 5 Percent of Families (All Races): 1947 to 2010,” and Table F-5, “Race and Hispanic Origin of Householder—Families by Median and Mean Income: 1947 to 2010.” Data are inflated to 2011 dollars using the CPI-U-RS and then indexed to 1979=100.

<strong>Figure 2L. Income growth for families at the 20th, 50th, and 95th percentiles, by nativity, 1993–2010. </strong>Underlying data are from Current Population Survey Annual Social and Economic Supplement microdata; see Appendix A for details. Data are inflated to 2011 dollars using the CPI-U-RS and then indexed to 1993=100.

<strong>Figure 2M. Change in real annual household income, by income group, 1979–2007.</strong> Underlying data are from the Congressional Budget Office, <em>Average Federal Taxes by Income Group</em>, “Sources of Income for All Households, by Household Income Category, 1979 to 2007” [Excel spreadsheet]. Cumulative growth is calculated by dividing the average pretax income in the base year (1979) into average pretax income in each subsequent year (1980–2007). The data provide average pretax income for the bottom, second, middle, fourth, and top fifths, and for the top 10, 5, and 1 percents. For the 80th–&lt;90th percentile, average pretax income is calculated by subtracting the aggregate income of the top 10 percent from aggregate income of the top fifth and dividing by the total number of households in the 80th–&lt;90th percentile. Aggregate income is calculated by multiplying the number of households in each income group by average pretax income. The number of households is calculated by subtracting the number of households in the top 10 percent from the number of households in the top fifth. This same procedure is done between the top 10 percent and top 5 percent to calculate average pretax income for the 90th–&lt;95th percentile and between the top 5 percent and top 1 percent to calculate the average pretax income for the 95th–&lt;99th percentile. Note that this publicly available CBO dataset is rounded, and produces slightly different income dollar values than the unpublished, unrounded CBO data underlying tables 2.13 and 2.16. Data are inflated to 2011 dollars using the CPI-U-RS, and then indexed to 1979=0.

<strong>Text Box Figure 2AA. Share of income held by high-income groups, 1913–2010. </strong>Underlying data are from Piketty and Saez (2012, Table A-3), Current Population Survey Annual Social and Economic Supplement <em>Historical Income Tables</em>, Table H-2, “Share of Aggregate Income Received by Each Fifth and Top 5 Percent of Households,” and the Congressional Budget Office <em>Average Federal Taxes by Income Group </em>report, “Average Pre-Tax Income for All Households, by Household Income Category, 1979–2007” [Excel spreadsheet]. The top 5 percent share is shown because the CPS data do not allow examination of the top 1 percent.

<strong>Text Box Figure 2AB. Share of income held by top 1 percent in developed countries, 1913–2009.</strong> Underlying data are from the World Top Incomes Database.

<strong>Figure 2N. Change in the share of market income and post-tax, post-transfer income that households claim, by income group, 1979–2007.</strong> Underlying data are from the Congressional Budget Office, <em>Average Federal Taxes by Income Group</em>, “Pre-Tax Income Shares All Households, by Household Income Category, 1979–2007,” “After-Tax Income Shares for All Households, by Household Income Category, 1979–2007,” and “Sources of Income for All Households, by Household Income Category, 1979 to 2007” [Excel spreadsheets] and unpublished health benefit data pertaining to this report. The shares of pre- and post-tax income are taken directly from the first two datasets cited here. The change in market income is then expressed as a share of the overall change in pretax income (transfers are essentially the only nonmarket income type that changes the pretax income shares).

<strong>Figure 2O. Effect of tax policies on each household income group’s share of total income, 1979 and 2007, and the difference needed in 2007 to preserve 1979 post-tax shares.</strong> Underlying data are from the Congressional Budget Office,
<em>Average Federal Taxes by Income Group</em>, “Average Federal Tax Rates for All Households, by Comprehensive Household Income Quintile, 1979–2007”and “Sources of Income for All Households, by Household Income Category, 1979 to 2007” [Excel Spreadsheets].

<strong>Figure 2P. Average effective federal tax rates, by household income group, 1979–2007.</strong> Underlying data are from the Congressional Budget Office, <em>Average Federal Taxes by Income Group</em>, “Average Federal Tax Rates for All Households, by Comprehensive Household Income Quintile, 1979–2007” [Excel spreadsheet] and “Effective Federal Tax Rates for All Households, by Comprehensive Household Income Category, 1979 to 2005 (Percent)” [Excel spreadsheet supplement to <em>Historical Effective Federal Tax Rates: 1979 to 2005</em>]. The tax rates for the top .01, top 0.1 and top 1.0 percent are given by CBO. The tax rates for the 20th–&lt;90th percentile, 90th–&lt;95th percentile, and the 95th–&lt;99th percentile are calculated by taking an average of each income groups’ tax rate weighted by their share of total income.

<strong>Figure 2Q. Average effective federal tax rates, by income group, 1960–2004.</strong> Underlying data are from Piketty and Saez (2007), Table 2, “Federal Rates by Income Groups, 1960 to 2004.” The top .01 percent, the 99.9th–&lt;99.99th percentile, 99.5th–&lt;99.9th percentile, and 99.0–&lt;99.5th percentile data are provided. The 20th–&lt;99th percentile tax rate was calculated as an average of each income groups’ tax rate weighted by their share of total income.

<strong>Figure 2R. Change in real cash and medical transfer income, by income group, 1979–2007.</strong> Underlying data are from the Congressional Budget Office, <em>Average Federal Taxes by Income Group</em>, “Sources of Income for All Households, by Household Income Category, 1979 to 2007” [Excel spreadsheet] and unpublished data related to the same report on the composition of in-kind income, with a breakout for health spending (both government transfers as well as employer-sponsored insurance benefits).

<strong>Figure 2S. Change in tax rate, transfer rate, and tax rate net of transfers, by income group, 1979–2007.</strong> Data in Figure 2S are a subset of the data in Table 2.8.

<strong>Figure 2T. Change in real annual household wages, by income group, 1979–2007.</strong> Underlying data are from the Congressional Budget Office <em>Average Federal Taxes by Income Group</em>, “Sources of Income for All Households, by Household Income Category, 1979 to 2007” [Excel spreadsheet]. Cumulative growth is calculated by dividing the average wages in the base year (1979) into average wages in each subsequent year (1980–2007). Average wages by income group are calculated by multiplying the share of wages by the average pretax income in each income group. See Figure 2M notes for calculations of the 80th–&lt;90th percentile, 90th–&lt;95th percentile, and 95th–&lt;99th percentile. Data are inflated to 2011 dollars using the CPI-U-RS, and then indexed to 1979=0.

<strong>Figure 2U. Change in real household capital income, by income group, 1979–2007.</strong> Underlying data are from the Congressional Budget Office, <em>Average Federal Taxes by Income Group</em>, “Sources of Income for All Households, by Household Income Category, 1979–2007” [Excel spreadsheet]. Cumulative growth is calculated by dividing the average capital income in the base year (1979) into average capital income in each subsequent year (1980–2007). Average capital income by income group is calculated by multiplying the share of capital income by the average pretax income in each income group. See Figure 2M notes for calculations of the 80th–&lt;90th percentile, 90th–&lt;95th percentile, and 95th–&lt;99th percentile; see Table 2.4 notes for explanation of capital income. Data are inflated to 2011 dollars using the CPI-U-RS, and then indexed to 1979=0.

<strong>Figure 2V. Share of total household capital income claimed, by income group, 1979–2007.</strong> Underlying data are from the Congressional Budget Office, <em>Average Federal Taxes by Income Group</em>, “Sources of Income for All Households, by Household Income Category, 1979 to 2007” [Excel spreadsheet]. The share of capital income is each income group’s capital income share of the total capital income for all income groups. See Table 2.4 notes for the calculations for income group breakdowns and definition of capital income.

<strong>Figure 2W. Pretax and post-tax profit rates, 1959–2010.</strong> Underlying data are from the Bureau of Economic Analysis National Income and Product Accounts tables, Table 1.14, “Gross Value Added of Domestic Corporate Businesses in Current Dollars and Gross Value added of Nonfinancial Domestic Corporate Business in Current and Chained Dollars” and Fixed Assets Accounts tables, Table 6.1, “Current-Cost Net Stock of Private Fixed Assets by Industry Group and Legal Form of Organization.” For calculations of pre- and post-tax profit rate, see Table 2.12 notes.

<strong>Figure 2X. Capital share of total corporate-sector income, actual and counterfactual holding 1979 profit rate constant, 1979–2010.</strong> Underlying data are from the Bureau of Economic Analysis, National Income and Product Accounts tables, Table 1.14, “Gross Value Added of Domestic Corporate Businesses in Current Dollars and Gross Value added of Nonfinancial Domestic Corporate Business in Current and Chained Dollars” and Fixed Assets Accounts tables, Table 6.1, “Current-Cost Net Stock of Private Fixed Assets by Industry Group and Legal Form of Organization.” For calculations of pretax and post-tax profit rate, see Table 2.12 notes.

<strong>Figure 2Y. Share of total household income growth attributable to various income groups, 1979–2007.</strong> Underlying data are from the Congressional Budget Office <em>Average Federal Taxes by Income Group</em>, “Sources of Income for All Households, by Household Income Category, 1979 to 2007” [Excel spreadsheet]. Each group’s contribution to overall income growth is calculated by multiplying the change in its average income from 1979 to 2007 by its share of the distribution (where, for example, the share of the distribution for the top 1 percent is .01), and dividing the result by the change in overall average income growth over the same time period. For pretax income calculations of the 90th–&lt;95th percentile and 95th–99th percentile, see Figure 2M notes.

<strong>Figure 2Z. Change in household income, as reported by CBO comprehensive income data and CPS money income data, by income group, 1979–2007. </strong>Underlying data are from Congressional Budget Office <em>Average Federal Taxes by Income Group </em>report, “Sources of Income for All Households, by Household Income Category, 1979 to 2007” [Excel spreadsheet], and Current Population Survey Annual Social and Economic Supplement <em>Historical Income Tables,</em> Table F-3, “Mean Income received by each fifth and top 5 percent of all families, 1966–2010.” Percentage change of household income is calculated between the years 1979 and 2007. Note that this publicly available CBO dataset is rounded, and produces slightly different income dollar values than the unpublished, unrounded CBO data underlying tables 2.13 and 2.16. Data are inflated to 2011 dollars using the CPI-U-RS, and then indexed to 1979=0.

<strong>Figure 3A. Median family income over the householder’s working life, by birth cohort.</strong> Data are from Current Population Survey Annual Social and Economic Supplement <em>Historical Income Tables</em>, Table F-11, “Age of Householder—Families, All Races by Median and Mean Income: 1947 to 2010.” Data are inflated to 2011 dollars using the CPI-U-RS (Consumer Price Index Research Series Using Current Methods). Income measured is family money income, defined in Chapter 2.

<strong>Figure 3B. Share of families in the bottom and top income fifths in 1994 ending up in various income fifths in 2004.</strong> Data are from Acs and Zimmerman (2008a), Table 2, “Quintile Transitions, Two-Year Average Income (Relative Mobility).” Data for other years and relationships are available on <em>The State of Working America</em> website (http://stateofworkingamerica.org/).

<strong>Figure 3C. Share of workers with large shifts in real annual earnings from 2002 to 2003, by earnings fifth.</strong> Data are from Dahl and Schwabish (2008), Table 1, “Distribution of Changes in Workers’ Annual Real Earnings from 2002–2003, by Sex, Age, and Earnings Quintile.” The sample consists of workers age 25 to 55 who had earnings from employment covered by Social Security in 2002 or 2003. Earnings include wages and salaries, tips, and other forms of compensation; they exclude self-employment income and deferred compensation. Before the percentage change was calculated, earnings were adjusted for inflation using the CPI-U-RS.

<strong>Figure 3D. Share of taxpayers at the top of the income distribution in 1996 ending up in various income groups in 2005.</strong> Data are from U.S. Department of the Treasury (2007), Table 2, “The Degree of Mobility Remains Substantial after Restricting the Analysis to Taxpayers Included in the Panel of Tax Returns.” The table uses the tax returns of primary and secondary nondependent taxpayers who were age 25 and older in 1996 and filed for both 1996 and 2005. Income cutoffs for the percentiles are based only on the tax returns of the panel population. Income is cash income in 2005 dollars.

<strong>Figure 3E. Characteristics associated with leaving the bottom income fifth.</strong> The figure is adapted from Acs and Zimmerman (2008b), Figure 5, “Characteristics Associated with Leaving the Bottom Quintile.” Coefficients are based on a linear probability regression that includes these characteristics as well as dummy variables for age, the presence of children, and the presence of other adults in the household. <em>Own</em> and <em>spouse work hours</em> are measured in thousand-hour units. Acs and Zimmerman do not differentiate between spouses and permanent cohabiters, and interact the spouse hours variable with a dummy variable for the spouse’s presence. Only characteristics with statistically significant coefficients in at least one time period are shown. In the 1984–1994 time period, the coefficients for <em>white</em>, <em>more than high school</em>, <em>disability</em>, and <em>spouse present</em> are statistically significant at the 99 percent confidence level; <em>high school education</em> is statistically significant at the 95 percent confidence level; and <em>male</em>, <em>homeowner</em>, <em>own hours</em>, and <em>spouse’s hours</em> are statistically significant at the 90 percent confidence level. In the 1994–2004 time period, the coefficients for <em>more than high school</em> and <em>own hours</em> are statistically significant at the 99 percent confidence level.

<strong>Figure 3F. Characteristics associated with entering the bottom income fifth.</strong> The figure is adapted from Acs and Zimmerman (2008b), Figure 6, “Characteristics Associated with Entering the Bottom Quintile.” Coefficients are based on a linear probability regression that includes these characteristics as well as dummy variables for age, education, the presence of children, and own work hours. <em>Own</em> and <em>spouse work hours</em> are measured in thousand-hour units. Acs and Zimmerman do not differentiate between spouses and permanent cohabiters, and interact the spouse hours variable with a dummy variable for the spouse’s presence. Only characteristics with statistically significant coefficients in at least one time period are shown. In the 1984–1994 time period, the coefficients for <em>middle fifth</em>, <em>fourth fifth</em>, and <em>top fifth</em> are statistically significant at the 99 percent confidence level; <em>male</em> and <em>spouse present</em> are statistically significant at the 95 percent confidence level; and <em>spouse work hours</em> is statistically significant at the 90 percent confidence level. In the 1994–2004 time period, the coefficients for <em>disability</em>, <em>fourth fifth</em>, and <em>top fifth</em> are statistically significant at the 99 percent confidence level; <em>white</em> is statistically significant at the 95 percent confidence level; and <em>homeowner</em>, <em>other adult present</em>, and <em>middle fifth</em> are statistically significant at the 90 percent confidence interval.

<strong>Figure 3G. Likelihood that sons of low-earning fathers end up above various earnings thresholds as adults, depending on estimated ease of mobility.</strong> Data are from Solon (1989), Table 5, “Probability that Son’s Long-Run Status Is in Specified Decile Given Percentile of Father’s Status.” Data are from the 1985 follow-up to the 1968 Panel Study of Income Dynamics. “Earnings” refers to wages.

<strong>Figure 3H. Intergenerational correlations between the earnings of fathers and sons in OECD countries.</strong> The figure is adapted from Corak (2011), Figure 1, “Comparable Estimates of the Intergenerational Elasticity between Father and Son Earnings for the United States and Twenty Four Other Countries.” “Earnings” refers to wages.

<strong>Figure 3I. Share of sons of fathers in the bottom earnings fifth ending up in the bottom or top two-fifths as adults, by country.</strong> Data are from Jantti et al. (2006), Table 12, “Intergenerational Mobility Tables—Earnings Quintile Group Transition Matrices Corrected for Age for Fathers and Sons.” These results include only those father-son pairs that have non-zero earnings (wages).

<strong>Figure 3J. Share of daughters of fathers in the bottom earnings fifth ending up in the bottom or top two-fifths as adults, by country.</strong> Data are from Jantti et al. (2006), Table 13, “Intergenerational Mobility Tables—Earnings Quintile Group Transition Matrices Corrected for Age for Fathers and Daughters.” These results include only those father-daughter pairs that have non-zero earnings (wages).

<strong>Figure 3K. Share of children in the bottom income fourth ending up in either the bottom or top income fourth as adults, by race.</strong> Data are from Hertz (2006), Table 1, “Mobility Experience of Children Born in the Bottom Quartile, By Race.” The quartile boundaries change over time, as real incomes grow. The black-white gap in the likelihood of upward mobility was statistically significant at the 1 percent level, and persists after controlling for one’s starting position within the quartile, and for parental education.

<strong>Figure 3L. Share of children from various earnings fifths ending up in the bottom fifth as adults, by race.</strong> Data are from Mazumder (2011), Table 7, “Transition Matrices by Race Using SIPP-SSA Sample.” Both panels use subsamples drawn from a sample of 16,782 men from the Survey of Income and Program Participation and Social Security Administration data and use a multiyear average of sons’ earnings over 2003–2007 and parents’ earnings over 1978–1986.

<strong>Figure 3M. Share of children in the bottom and top wealth fifths ending up in various wealth fifths as adults.</strong> Data are from Charles and Hurst (2002), Table 2, “Intergenerational Transition Matrix of Age-Adjusted Log Wealth Position.” The sample includes all PSID parent-child pairs in which the following conditions were met (1,491 pairs): Parents were in the survey in 1984–1989 and alive in 1989, the child was in the survey in 1999, the parent was not retired and was between age 25 and 65 in 1984, the child was between age 25 and 65 in 1999, and the child and parent both had positive wealth when measured.

<strong>Figure 3N. Share of entering classes at top universities and community colleges coming from families in various socioeconomic fourths.</strong> Data are from Carnevale and Rose (2003), Table 1.1, “Socioeconomic Status of Entering Classes.” Socioeconomic status is measured by a composite score that includes family income, parental education, and parental occupation.

<strong>Figure 3O. Share of students completing college, by socioeconomic status and eighth-grade test scores.</strong> Data are from Fox, Connolly, and Snyder (2005), Table 21, “Percentage Distribution of 1988 Eighth-Graders’ Educational Attainment by 2000, by Eighth-Grade Mathematics Achievement and Selected Student Characteristics: 2000.” Socioeconomic status is measured by a composite score that includes family income, parental education, and parental occupation.

<strong>Figure 3P. Share of adults remaining in the same income fifth they were in as children, by college attainment.</strong> Data are from Isaacs, Sawhill, and Haskins (2008), Figure 6, “Chances of Getting Ahead for Children with and without a College Degree, from Families of Varying Income.”

<strong>Figure 3Q. Intergenerational mobility and income inequality in 22 countries.</strong> The figure is adapted from Corak (2012), Figure 2, “More Inequality at a Point in Time Is Associated with Less Generational Earnings Mobility in Twenty Five Countries with Comparable Estimates of the Intergenerational Elasticity Between Father and Son Earnings.” Note that data points for Italy and the United Kingdom overlap, and that the upward sloping line is the least squares fitted regression line.

<strong>Figure 3R. Distance between income groups in the United States versus the European Union (hypothetical).</strong> Authors’ illustration.

<strong>Figure 3S. Share of people in the bottom and top family income fifths moving along the income scale, 1970–1980 to 1995–2005.</strong> The figure is adapted from Bradbury (2011), Figure 2, “Position-relative Origin-specific Mobility for Poorest and Richest Quintiles.”

<strong>Figure 3T. Share of working-age individuals experiencing a 50% or greater drop in family income over two years, 1971–2004.</strong> Data are from Hacker and Jacobs (2008), Figure C, “Prevalence of a 50% or Greater Drop in Family Income.” The line traces the share of individuals age 25 to 61 who experience a 50 percent or greater drop in before-tax total family income (adjusted for family size) from one year to two years later. Data after 1996 are only available every two years.

<strong>Figure 3U. Elasticities between parental income and sons’ earnings, 1950–2000.</strong> Data are from Aaronson and Mazumder (2007), Table 1,“Estimates of the IGE Using Census IPUMS Data.” Data reflect annual family income for the parents and annual earnings for the sons.

<strong>Figure 3V. Share of 25-year-olds from each family income fourth without a college degree, by birth cohort.</strong> Data are from Bailey and Dynarski (2011), Figure 3, “Fraction of Students Completing College, by Income Quartile and Year of Birth,” which is based on data from the National Longitudinal Survey of Youth, 1979 and 1997. Family income fourths are those of 25-year-olds when they were children.

<strong>Table 4.1. Average wages and work hours, 1967–2010.</strong>Productivity data, which measure output per hour of the total economy, including the private and public sectors, are from an unpublished series available from the Bureau of Labor Statistics Labor Productivity and Costs program on request. The wage-level data are based on the authors’ tabulations of Current Population Survey Annual Social and Economic Supplement (CPS-ASEC, also known as the March CPS) microdata files using a series on annual, weekly, and hourly wages for wage and salary workers. See Appendix B for the sample definition and other information. The weekly and hourly wage data are “hour weighted,” obtained by dividing annual wages by weeks worked and annual hours worked. The 1967 and 1973 values are derived from unpublished tabulations provided by Kevin Murphy from an update of Murphy and Welch (1989); they include self-employment as well as wage and salary workers. The values displayed in this table were bridged from CPS 1979 values using the growth rates in the Murphy and Welch series. Hours of work were derived from differences between annual, weekly, and hourly wage trends.

<strong>Table 4.2. Average hourly pay and pay inequality, 1948–2011.</strong>The data in the top panel are computed from the Bureau of Economic Analysis National Income and Product Accounts (NIPA) tables. “Wages and salaries” are calculated by dividing wage and salary accruals (NIPA Table 6.3) by hours worked by full-time and part-time employees (NIPA Table 6.9). “Total compensation” is the sum of wages and salaries and benefits (it includes payroll taxes and health, pension, and other nonwage benefits). Payroll taxes are calculated as total compensation (NIPA Table 6.2) minus the sum of volunteer benefits (sum of health and nonhealth benefits; see NIPA Table 6.11) and wages and salaries. “Benefits” is the difference between total compensation and wages and salaries. These data were deflated using the NIPA personal consumption expenditure (PCE, chain-weighted) index, with health insurance adjusted by the PCE medical care (chained) index. These data include both public- and private-sector workers.

The data in the Employer Costs for Employee Compensation (ECEC) panel come from the BLS National Compensation Survey’s employment cost trends and benefits data and provide cost levels for March for private-sector workers, available starting in 1987. We categorize wages and salaries differently than BLS, putting all wage-related items (including paid leave and supplemental pay) into the hourly wage/salary column. This makes the definition of wages and salaries comparable to workers’ W-2 earnings and to the definition of wages in the CPS Outgoing Rotation Group (ORG) data that are tabulated for other tables in this chapter. Benefits, in our definition, include only payroll taxes, pensions, insurance, and “other” benefits. The sum of wages and salaries and benefits makes up total compensation. It is important to use the ECEC (the current-weighted series) rather than the other series from the same National Compensation Survey (NCS) data, the ECI (the fixed-weighted series), because composition shifts (in the distribution of employment across occupations and industries) can have large effects over time. Employer costs for insurance are deflated by the medical-care component of the CPI-U-RS (Consumer Price Index Research Series Using Current Methods). All other pay is deflated by the CPI-U-RS for “all items.” Inflation is measured for the first quarter of each year. Wage and compensation inequality measures are drawn from Pierce (2010). Pierce computes these from the NCS microdata, the data used to calculate the ECI and ECEC data.

<strong>Table 4.3. Hourly and weekly earnings of private production and nonsupervisory workers, 1947–2011.</strong> Underlying data are from the Bureau of Labor Statistics Current Employment Statistics program data from the <em>Employment, Hours, and Earnings–National </em>database, deflated using CPI-U-RS.

<strong>Table 4.4. Hourly wages of all workers, by wage percentile, 1973–2011. </strong>Table is based on analysis of CPS wage data described in Appendix B.

<strong>Table 4.5. Hourly wages of men, by wage percentile, 1973–2011. </strong>Table is based on analysis of CPS wage data described in Appendix B.

<strong>Table 4.6. Hourly wages of women, by wage percentile, 1973–2011.</strong> Table is based on analysis of CPS wage data described in Appendix B.

<strong>Table 4.7. Change in wage groups’ shares of total wages, 1979–2010.</strong> Data are taken from Kopczuk, Saez, and Song (2010), Table A-3. Data for 2006 through 2010 are extrapolated from 2004 data using changes in wage shares computed from Social Security Administration wage statistics (data for 2010 at http://www.ssa.gov/cgi-bin/netcomp.cgi). The final results of the paper by Kopczuk, Saez, and Song printed in a journal used a more restrictive definition of wages so we employ the original definition, as recommended in private correspondence with Kopczuk. SSA provides data on share of total wages and employment in annual wage brackets such as for those earning between $95,000.00 and $99,999.99. We employ the midpoint of the bracket to compute total wage income in each bracket and sum all brackets. Our estimate of total wage income using this method replicates the total wage income presented by SSA with a difference of less than 0.1 percent. We use interpolation to derive cutoffs building from the bottom up to obtain the 0–90th percentile bracket and then estimate the remaining categories. This allows us to estimate the wage shares for upper wage groups. We use these wage shares computed for 2004 and later years to extend the Kopczuk, Saez, and Song series by adding the changes in share between 2004 and the relevant year to their series. To obtain absolute wage trends we use the SSA data on the total wage pool and employment and compute the real wage per worker (based on their share of wages and employment) in the different groups in 2011 dollars.

<strong>Table 4.8. Change in annual wages, by wage group, 1979–2010.</strong> See note to Table 4.7.

<strong>Table 4.9. Specific fringe benefits, 1987–2011. </strong>Table is based on ECEC data described in note to Table 4.2.

<strong>Table 4.10. Employer-provided health insurance coverage, by demographic and wage group, 1979–2010.</strong> Table is based on tabulations of CPS-ASEC data samples of private wage-and-salary earners ages 18–64 who worked at least 20 hours per week and 26 weeks per year. This sample is chosen to focus on those with regular employment. Coverage is defined as being included in an employer-provided plan for which the employer paid for at least some of the coverage. As with other CPS microdata analyses presented in the book, race/ethnicity categories are mutually exclusive (i.e., white non-Hispanic, black non-Hispanic, and Hispanic any race).

<strong>Table 4.11. Employer-provided pension coverage, by demographic and wage group, 1979–2010. </strong>Table is based on CPS-ASEC data on pension coverage, using the sample described in the note to Table 4.10. As with other CPS microdata analyses presented in the book, race/ethnicity categories are mutually exclusive (i.e., white non-Hispanic, black non-Hispanic, and Hispanic any race).

<strong>Table 4.12. Share of workers with paid leave, by wage group, 2011.</strong>Computed from BLS Employee Benefits Survey, <em>Holiday, Vacation, Sick, and Other Leave Benefits, March 2011,</em> data tables 34, 36, and 38; http://www.bls.gov/ncs/ebs/benefits/2011/benefits_leave.htm.

<strong>Table 4.13. Dimensions of wage inequality, by gender, 1973–2011.</strong> All of the data are based on analyses of the CPS-ORG data described in Appendix B and used in various tables. The measures of “total wage inequality” are natural logs of wage ratios (multiplied by 100) computed from Tables 4.5 and 4.6. The exception is 1979 data for women, which are 1978–1980 averages; we use these to smooth the volatility of the series, especially at the 10th percentile. The “between-group inequalities” are computed from regressions of the log of hourly wages on education categorical variables (advanced, college only, some college, less than high school with high school omitted), experience as a quartic, marital status, race, and region (4). The college/high school and high school/less-than-high-school premiums are simply the coefficient on “college” and “less than high school” (expressed as the advantage of “high school” over “less than high school” wages). The experience differentials are the differences in the value of age (calculated from the coefficients of the quartic specification) evaluated at 25, 35, and 50 years old. “Within-group wage inequality” is measured as the root mean square error from the same log wage regressions used to compute age and education differentials.

<strong>Table 4.14. Hourly wages by education, 1973–2011. </strong>Table is based on tabulations of CPS wage data described in Appendix B. See Appendix B for details on how a consistent measure of education was developed to bridge the change in coding in 1992.

<strong>Table 4.15. Hourly wages of men, by education, 1973–2011. </strong>See note to Table 4.14.

<strong>Table 4.16. Hourly wages of women, by education, 1973–2011. </strong>See note to Table 4.14.

<strong>Table 4.17. Educational attainment of the employed, by gender and nativity, 2011. </strong>Table is based on analysis of CPS wage earners. The data are described in Appendix B. The categories are as follows: “less than high school” is grade 1–12 or no diploma; “high school/GED” is high school graduate diploma or equivalent; “some college” is some college but no degree; “associate degree” is occupational or academic associate degree; “college degree” is a bachelor’s degree; and “advanced degree” is a master’s, professional, or doctoral degree.

<strong>Table 4.18. Hourly wages of entry-level and experienced workers, by gender and education, 1973–2011.</strong> Table is based on analysis of CPS wage data described in Appendix B. Entry-level wages are measured for a seven-year window starting a year after normal graduation, which translates to ages 19–25 for high school graduates and ages 23–29 for college graduates.

<strong>Table 4.19. Hourly wages by wage percentile, gender, and education, 1973–2011.</strong> Table is based on analysis of CPS wage data described in Appendix B.

<strong>Table 4.20. Contribution of within-group and between-group inequality to total wage inequality, 1973–2011.</strong> Data are from the CPS-ORG sample described in Appendix B. “Overall wage inequality” is measured as the standard deviation of log wages. “Within-group wage inequality” is the mean square error from log wage regressions (the same ones used for Table 4.13). “Between-group wage inequality” is the difference between the overall and within-group wage inequalities and reflects changes in all of the included variables: education, age, marital status, race, ethnicity, and region.

<strong>Table 4.21. Hourly wage growth by gender and race/ethnicity, 1989–2011.</strong> Table is based on analysis of CPS wage data described in Appendix B. As with other CPS microdata analyses presented in the book, race/ethnicity categories are mutually exclusive (i.e., white non-Hispanic, black non-Hispanic, and Hispanic any race).

<strong>Table 4.22. Gender wage gap, 1973–2011.</strong> Wages and ratios are based on 50th-percentile wages from Tables 4.5 and 4.6 (CPS-ORG data).

<strong>Table 4.23. Factors contributing to the productivity/compensation gap, 1973–2011. </strong>Table is based on analysis of Mishel and Gee (2012), Table 1. Mishel and Gee present a decomposition of the gap between productivity and median hourly compensation. This has been reconfigured to eliminate the gap between median hourly wages and compensation so the decomposition is between productivity and median hourly compensation.

<strong>Table 4.24. Impact of rising and falling unemployment on wage levels and gaps, 1979–2011. </strong>Table is based on analyses of yearly wage decile data from Tables 4.5 and 4.6 (see Appendix B), and of unemployment data using model from Katz and Krueger (1999). The unemployment rate is from the Current Population Survey. The simulated effect of change of unemployment presented in the table was calculated by regressing the log-change of nominal wages on the lagged log-change of the CPI-U-RS (but, following Katz and Krueger [1999], the coefficient is constrained to equal 1), the unemployment rate, lagged productivity growth, and dummies for various periods (1989–1995, 1996–2000, 2001–2007). Using these models, wages were predicted for the periods in the table given a simulated unemployment rate series in which unemployment remains fixed at its starting-year level. So in the 1979 to 1985 period, unemployment was fixed at its 1979 level and not allowed to rise (as actually happened) throughout the period. The “estimated cumulative impact of unemployment” shows the difference between actual wages and the wages when unemployment was held fixed in the starting year.

<strong>Table 4.25. Annual pay in expanding and contracting industries, 1979–2007.</strong> These data reflect the average (annual) wages, benefits, and compensation of the net new employment in each period based on changes in industry composition. The employment data are payroll counts from the BLS Current Employment Statistics, and the pay data are from 2008 Bureau of Economic Analysis NIPA tables (calculated per payroll employee). The pay of the net new employment is a weighted average of the pay by industry in which the weights are the changes in each industry’s employment share over the period.

<strong>Table 4.26. Employer health care costs as a share of wages, 1948–2010.</strong> Table is based on analysis of National Income and Product Accounts data. Wage data are from NIPA Table 6.3, and group health insurance data are from NIPA Tables 6.11A-C, and 6.11D.

<strong>Table 4.27. Employer health care costs as a share of wages, by wage fifth, 1996–2008. </strong>Table is based on analysis of Burtless and Milusheva (2012) based on Medical Expenditure Panel Survey. The authors provide data by decile which we aggregated to fifths. The premiums include both those enrolled and not enrolled in employer plans. The premiums were estimated by Burtless and Milusheva using various imputation methods.

<strong>Table 4.28. Impact of trade balance in manufacturing on employment and wages, by education, 1979–2005. </strong>Table is based on analysis of Bivens (2008).

<strong>Table 4.29. Impact of trade with low-wage countries on college/noncollege wage gap, 1973–2011. </strong>Table is an update of Bivens’s (2008) reanalysis of Krugman (1995) using 2011 data.

<strong>Table 4.30. Characteristics of offshorable and non-offshorable jobs. </strong>Table reflects authors’ analysis of the Bernstein, Lin, and Mishel (2007) analysis of data of Blinder (2007), matching Blinder’s occupational codes to the BLS Occupational Employment Statistics (OES) survey (http://www.bls.gov/oes/) and Blinder and Krueger (2009) Table 4.

<strong>Table 4.31. Mexican and other immigrants’ share of U.S. workforce, by gender, 1940–2011. </strong>Data are from Figure 1 in Borjas and Katz (2005) and authors’ computations of Current Population Survey basic monthly microdata for 2000 and 2011.

<strong>Table 4.32. Educational attainment of immigrants, by gender, 1940–2011.</strong>Data are from Table 2 in Borjas and Katz (2005) and authors’ computations of Current Population Survey basic monthly microdata for 2000 and 2011.

<strong>Table 4.33. Union wage premium by demographic group, 2011.</strong> “Percent union” is tabulated from CPS-ORG data (see Appendix B) and includes all those covered by unions. “Union premium” values are the coefficients on union in a model of log hourly wages with controls for education, experience as a quartic, marital status, region, industry (12) and occupation (9), race/ethnicity, and gender where appropriate. For this analysis we only use observations that do not have imputed wages because the imputation process does not take union status into account and therefore biases the union premium toward zero. See Mishel and Walters (2003). As with other CPS microdata analyses presented in the book, race/ethnicity categories are mutually exclusive (i.e., white non-Hispanic, black non-Hispanic, and Hispanic any race).

<strong>Table 4.34. Union premiums for health, retirement, and paid leave benefits.</strong> Table is based on Table 4 in Mishel and Walters (2003), which draws on Buchmueller, DiNardo, and Valletta (2001).

<strong>Table 4.35. Union impact on paid leave, pension, and health benefits.</strong> Table is based on Table 3 in Mishel and Walters (2003), which draws on Pierce (1999), Tables 4, 5, and 6.

<strong>Table 4.36. Effect of union decline on male wage differentials, 1978–2011. </strong>This analysis replicates, updates, and expands on Freeman (1991), Table 2, using the CPS-ORG sample used in other analyses (see Appendix B). The year 1978, rather than 1979, is the earliest year analyzed because we have no union membership data in our 1979 sample. “Percent union” is the share covered by collective bargaining. The “union wage premium” for a group is based on the coefficient on collective bargaining coverage in a regression of hourly wages on a simple human capital model (the same one used for estimating education differentials, as described in note to Table 4.13), with major industry (12) and occupation (9) controls in a sample for that group. The change in union premium across years, therefore, holds industry and occupation composition constant. Freeman’s analysis assumed the union premium was unchanged over time. We allow the union premium to differ across years so changes in the “union effect” on wages (the union wage premium times union coverage) are driven by changes in the unionization rate and the union wage premium. The analysis divides the percentage-point change in the union effect on wage differentials by the actual percentage-point change in wage differentials (regression-adjusted with simple human capital controls plus controls for other education or occupation groups) to determine the deunionization contribution to the change in the wage gaps among men, which, as a negative percent, indicates contribution to the growth of the wage gaps.

<strong>Table 4.37. Union wage premium for subgroups.</strong> The analysis builds on Mishel and Walters (2003), Table 2.3A and Gundersen (2003), Table 5.1 and Appendix C. Premium estimates by fifth are from Schmitt (2008); Card, Lemieux, and Riddle (2002); and Gittleman and Pierce (2007). Union coverage by fifth is from Schmitt (2008).

<strong>Table 4.38. Impact of deunionization on wage inequality, 1973–2007. </strong>Table is based on analysis of Western and Rosenfeld (2011), Table 2.

<strong>Table 4.39. Value of the minimum wage, 1960–2011.</strong> Data, deflated using CPI-U-RS, are from the U.S. Department of Labor Wage and Hour Division (2009); http://www.dol.gov/whd/minwage/chart.htm.

<strong>Table 4.40. Characteristics of workers affected by proposed minimum-wage increase to $9.80 in 2014.</strong> Table is based on Cooper (2012) analysis of CPS Outgoing Rotation Group microdata. As with other CPS microdata analyses presented in the book, race/ethnicity categories are mutually exclusive (i.e., white non-Hispanic, black non-Hispanic, and Hispanic any race).

<strong>Table 4.41. Minimum-wage impact on 50/10 wage gap, 1979–2009. </strong>Analysis is of Autor, Manning, and Smith (2010), Table 5.

<strong>Table 4.42. Role of executives and financial sector in income growth of top 1.0% and top 0.1%, 1979–2005. </strong>Table is based on authors’ analysis of Bakija, Cole, and Heim (2012) Tables 4, 5, 6a, and 7a, using tables that include capital-gains income. The Bakija, Cole, and Heim paper tabulates IRS tax returns and exploits the information on the primary and secondary taxpayer occupation data provided there.

<strong>Table 4.43. CEO compensation and CEO-to-worker compensation ratio, 1965–2011. </strong>Complete details on the data used to compute CEO compensation trends and the CEO-to-worker compensation ratio can be found in Mishel and Sabadish (2012), <em>Methodology for Measuring CEO Compensation and the Ratio of CEO-to-Worker Compensation</em> at http://www.epi.org/publication/wp293-ceo-to-worker-pay-methodology. We use executive compensation data from the ExecuComp database of Compustat, a division of Standard &amp; Poor’s. The ExecuComp database contains data on many forms of compensation for the top five executives at publicly traded U.S. companies in the S&amp;P 1500 Index for 1992–2010. We employ two definitions of annual CEO compensation based on different ways of measuring option awards. “Realized direct compensation,” referred to as “Options realized” in the table, is the sum of salary, bonus, restricted stock grants, options exercised, and long-term incentive payouts. It follows the definition of compensation used in previous editions of <em>The State of Working America</em>, which in turn adapted this definition from the <em>Wall Street Journal</em> (<em>WSJ</em>) annual report on CEO compensation (compensation reported by the <em>WSJ</em> has been compiled by various companies over the years, including Pearl Meyer, the Mercer Group, and the Hay Group and is the longest CEO pay series available to us). “Total direct compensation” (also a definition used in the <em>WSJ</em> series and labeled “Options granted” in the table) is the sum of salary, bonus, restricted stock grants, options granted (Compustat Black Scholes value), and long-term incentive payouts.

We define a CEO as an executive labeled a CEO by the variable CEOANN. Note that the executive flagged as the CEO may not necessarily be the highest-paid executive at the company. The CEOs included in our series are CEOs at the top 350 firms based on sales each year for 1992–2010.

Because no data for the compensation of an average worker in a firm exist, we create a proxy: the hourly compensation of a “typical” worker in a firm’s key industry. The wage measure is the production/nonsupervisory worker hourly earnings in that industry, the same series used in Table 4.3 for the entire private sector. We obtain compensation by multiplying the compensation wage ratio computed from NIPA Tables 6.3C and 6.3D. The hourly wages of production and nonsupervisory employees in 2011 were $19.47, 21 percent higher than the median hourly wage, so our proxy severely overstates the compensation of a typical worker and understates the CEO-to-worker pay ratio.

We use the growth in CEO compensation in the <em>WSJ</em> series to extend the CEO compensation series and the CEO-to-worker compensation ratio series backward. The <em>WSJ</em> series conducted by Pearl Meyer covered the years 1965, 1968, 1973, 1978, 1989, and 1992. We convert the compensation series to constant dollars using the CPI-U-RS and calculate the ratio of CEO compensation in each year as a fraction of the 1992 CEO compensation level. We then apply these ratios to the CEO compensation for 1992 calculated from the ExecuComp data. This moves the series backward in time so that the growth of CEO pay is the same as in the Pearl Meyer/<em>WSJ</em> series but is benchmarked to the levels in the ExecuComp series.

We make a similar set of computations to obtain a historical series for the CEO-to-worker compensation ratio. We start with the Pearl Meyer/<em>WSJ</em> series in constant dollars and divide it by an estimate of private-sector annual compensation of production/nonsupervisory workers in the same year. The compensation series is the real hourly compensation series presented in Figure 4B multiplied by 2,080 hours.

<strong>Table 4.44. Trends in education wage gaps, key wage group wage gaps, and relative supply of education, 1979–2011.</strong> The gross wage gap data are computed from underlying yearly data with selected years presented in Tables 4.4 and 4.8. The education wage gaps are computed from the same regressions for which results on college/high school and high school/less-than-high-school wage premiums are reported in Table 4.13, regressions of the log of hourly wages on education categorical variables (advanced degree, college only, some college, less than high school with high school omitted), experience as a quartic, marital status, race, and region (4). The college or more/noncollege differential is drawn from a similar regression except there is only one education dummy variable for those with a college degree or advanced degree. This estimate was also used in the analysis of trade’s impact on the college wage gap presented in Table 4.29.

<strong>Table 4.45. Inflation-adjusted hourly wage trends of college graduates, by occupation, 2000–2011. </strong>Table is based on tabulations of CPS-ORG data with a sample of those with a college degree (but no advanced degree). See Appendix B for information on the wage data.

<strong>Table 4.46. Effect of changing occupational composition on wages and on education and training requirements, 2010–2020.</strong> Table is based on analysis of Thiess (2012), Tables 5 and 6, and BLS Employment Projections Program (2012), Table 9.

<strong>Figure 4A. Cumulative change in total economy productivity and real hourly compensation of selected groups of workers, 1995–2011.</strong> Productivity data, which measure output per hour of the total economy, including private and public sectors, are from an unpublished series available from the Bureau of Labor Statistics Labor Productivity and Costs program on request. Wage measures are the annual data used to construct tables in this chapter: median hourly wages (at the 50th percentile) from Table 4.4 and hourly wages by education from Table 4.14. These are converted to hourly compensation by scaling by the real compensation/wage ratio from the Bureau of Economic Analysis National Income and Product Accounts (NIPA) data used in Table 4.2.

<strong>Figure 4B. Real hourly earnings and compensation of private production and nonsupervisory workers, 1947–2011.</strong> Wage data are from series used in Table 4.3. Wages are converted to hourly compensation by scaling by the real compensation/wage ratio from the NIPA data used in Table 4.2.

<strong>Figure 4C. Cumulative change in real hourly wages of men, by wage percentile, 1979–2011.</strong> See note to Table 4.5.

<strong>Figure 4D. Cumulative change in real hourly wages of women, by wage percentile, 1979–2011.</strong> See note to Table 4.6.

<strong>Figure 4E. Share of workers earning poverty-level wages, by gender, 1973–2011.</strong> Figure is based on analysis of Current Population Survey (CPS) wage data described in Appendix B. The poverty-level wage is calculated using an estimate of the four-person weighted average poverty threshold in 2011 of $23,010 (based on the 2010 threshold updated for inflation). This is divided by 2,080 hours to obtain a poverty-level wage of $11.06 in 2011. The poverty-level wage is roughly equal to two-thirds of the median hourly wage. This figure is deflated by CPI-U-RS (Consumer Price Index Research Series Using Current Methods) to obtain the poverty-level wage levels for other years. The threshold is available at the U.S. Census Bureau website.

<strong>Figure 4F. Share of workers earning poverty-level wages, by race and ethnicity, 1973–2011. </strong>As with other CPS microdata analyses presented in the book, race/ethnicity categories are mutually exclusive (i.e., white non-Hispanic, black non-Hispanic, and Hispanic any race).

<strong>Figure 4G. Share of total annual wages received by top earners, 1947–2010. </strong>See note to Table 4.7.

<strong>Figure 4H. Cumulative change in real annual wages, by wage group, 1979–2010. </strong>See note to Table 4.7.

<strong>Figure 4I. Share of private-sector workers with employer-provided health insurance, by race and ethnicity, 1979–2010.</strong> See note to Table 4.10.

<strong>Figure 4J. Share of pension participants in defined-contribution and defined-benefit plans, 1980–2004.</strong> Figure is based on Center for Retirement Research (2006), which used data from the Current Population Survey and the Department of Labor’s Annual Return/Report Form 5500 Series.

<strong>Figure 4K. Wage gaps among men, 1973–2011.</strong> Figure is based on ratios of yearly hourly wage by decile data presented in Table 4.5.

<strong>Figure 4L. Wage gaps among women, 1973–2011.</strong> Figure is based on ratios of yearly hourly wage by decile data presented in Table 4.6.

<strong>Figure 4M. Wage gap between the 95th and 50th percentiles, by gender, 1973–2011.</strong> Figure is based on ratios of yearly hourly wage by percentile data presented in Tables 4.5 and 4.6.

<strong>Figure 4N. College wage premium, by gender, 1973–2011. </strong>Differentials are estimated with controls for experience (as a quartic), region (4), marital status, race/ethnicity, and education, which are specified as dummy variables for less than high school, some college, college, and advanced degree. Log of hourly wage is the dependent variable. Estimates were made on the CPS-ORG data as described in Appendix B, and presented in Table 4.13.

<strong>Figure 4O. Share of the employed lacking a high school degree, by race/ethnicity and nativity status, 2011.</strong> Figure is based on tabulations of the full monthly CPS. See Appendix B for details on data.

<strong>Figure 4P. Real entry-level wages of high school graduates, by gender, 1973–2011. </strong>See note to Table 4.18.

<strong>Figure 4Q. Real entry-level wages of college graduates, by gender, 1973–2011.</strong> See note to Table 4.18.

<strong>Figure 4R. Share of recent high school graduates with employer health/pension coverage, 1979–2010.</strong> Data are computed from annual data series developed for Tables 4.10 and 4.11.The definition of recent high school graduates is the same as used in Table 4.18 for entry-level workers who are high school graduates; ages 19–25.

<strong>Figure 4S. Share of recent college graduates with employer health/pension coverage, 1979–2010. </strong>Data are computed from annual data series developed for Tables 4.10 and 4.11. The definition of recent college graduates is the same as used in Table 4.18 for entry-level workers who are college graduates; ages 23–29.

<strong>Figure 4T. Gender wage gap, by age cohort.</strong> See Moore and Shierholz (2007).

<strong>Figure 4U. Cumulative change in total economy productivity and real hourly compensation of production/nonsupervisory workers, 1948–2011.</strong> Productivity is based on unpublished Total Economy Productivity data from the Bureau of Labor Statistics Labor Productivity and Costs program. Hourly compensation for production/nonsupervisory workers is based on the wage data series used in Table 4.3. Wages are converted to hourly compensation by scaling by the real compensation/wage ratio from the NIPA data used in Table 4.2.

<strong>Figure 4V. Cumulative change in hourly productivity, real average hourly compensation, and median compensation, 1973–2011.</strong> Productivity and average hourly compensation are based on unpublished Total Economy Productivity data from the Bureau of Labor Statistics Labor Productivity and Costs program. Average hourly compensation includes those who are self-employed as well as wage and salary workers. See Mishel and Gee (2012) for more details. Median wages for all, men, and women are based on the data presented in Tables 4.4, 4.5, and 4.6, respectively. Wages are converted to hourly compensation by scaling by the real compensation/wage ratio from the NIPA data used in Table 4.2.

<strong>Figure 4W. Increase in worker wages from a 1 percentage-point fall in unemployment, by wage group.</strong> Estimates are based on a model employed by Katz and Krueger (1999). Annual changes in log wages are regressed on unemployment, lagged log-changes in the CPI-U-RS (but, following Katz and Krueger the coefficient on this is constrained to equal 1), lagged productivity growth, and dummies for 1989–1995, 1996–2000, and 2001–2007 (excluded period is 1979–1988). The sample covers the years 1979–2007.

<strong>Figure 4X. Employer health care costs as a share of annual wages, by wage fifth, 1996–2008.</strong> Figure is based on analysis of Burtless and Milusheva (2012), based on Medical Expenditure Panel Survey. See note to Table 4.27.

<strong>Figure 4Y. Imports, exports, and trade balance in goods as a share of U.S. GDP, 1947–2011.</strong> Figure is based on authors’ analysis of Bureau of Economic Analysis National Income and Product Accounts data.

<strong>Figure 4Z. Manufacturing imports as a share of U.S. GDP, 1973–2011.</strong> Figure is based on analysis of U.S. International Trade Commission Tariff and Trade data (series on manufacturing trade) and Bureau of Economic Analysis National Income and Product Accounts data on gross domestic product.

<strong>Figure 4AA. Relative productivity of U.S. trading partners, 1973–2011.</strong>Figure is based on analysis of United States International Trade Commission Tariff and Trade data and the Penn World Table (Heston, Summers, and Aten 2011). For each trading partner, their share of total imports was multiplied by their levels of GDP per worker relative to the United States (using data from the Penn World Tables). The resulting products were then summed to get the average productivity level of import trading partners. The same exercise was done for exports.

<strong>Figure 4AB. Wage premium of offshorable jobs, by gender and education. </strong>Figure is based on analysis of Bernstein, Lin, and Mishel (2007).

<strong>Figure 4AC. Union coverage rate in the United States, 1973–2011. </strong>Data are from Hirsch and Macpherson (2003), http://unionstats.gsu.edu/Hirsch-Macpherson_ILRR_CPS-Union-
Database.pdf; updated at unionstats.com. The data on union coverage begin in 1977 and are extended back to 1973, based on percentage-point changes in union membership shares in Hirsh and Macpherson (2003).

<strong>Figure 4AD. Real value of the minimum wage, 1960–2011.</strong> Underlying data are from U.S. Department of Labor Wage and Hour Division (2009), deflated using CPI-U-RS; see note to Table 4.39.

<strong>Figure 4AE. Minimum wage as a share of average hourly earnings, 1964–2011.</strong> The data are the minimum wage divided by the average hourly earnings of production and nonsupervisory workers. Minimum-wage levels are from Table 4.39, and average hourly earnings are from the series used in Table 4.3.

<strong>Figure 4AF. Real value of the federal minimum wage and share of workforce covered by higher state minimums, 1979–2011. </strong>Cooper (2012) update of Shierholz (2009).

<strong>Figure 4AG. Share of worker hours paid at or below the minimum wage, by gender, 1979–2009. </strong>Figure is based on analysis of Autor, Manning, and Smith (2010), Figure 1. Estimates are of the share of hours worked for reported wages equal to or less than the applicable state or federal minimum wage.

<strong>Figure 4AH. CEO-to-worker compensation ratio (options granted and options realized), 1965–2011. </strong>Figure is based on data developed for Table 4.43.

<strong>Figure 4AI. Growth in relative demand for college graduates, 1940–2005.</strong> Figure is based on authors’ analysis of Goldin and Katz (2008), Table 1.

<strong>Figure 4AJ. Cumulative change in real hourly wages of college graduates, by decile, 2000–2011. </strong>Figure is based on authors’ analysis of CPS-ORG data using a sample of college graduates (but no advanced degree). See Appendix B for data details<strong>. </strong>

<strong>Figure 4AK. Underemployment of college graduates, by age, 2000–2010. </strong>Figure is based on authors’ analysis of Fogg and Harrington (2011), Table 1. “Underemployment” occurs when a college graduate works in an occupation that does not require a college education.

<strong>Figure 4AL. Education needed in 2020 workforce and education levels of the 2011 workforce. </strong>Figure is based on authors’ analysis of Thiess (2012) for Table 4.46 and education attainment data from Table 4.17.

<strong>Table 5.1.</strong> <strong>Average annual change in employment, GDP, hours, and productivity, 1948–2011. </strong>Underlying data for total economy productivity are unpublished data provided to the authors by the Bureau of Labor Statistics Labor Productivity and Costs program.

<strong>Table 5.2. Labor force share and unemployment rate, by age, 1979–2011.</strong> Underlying data are from the Current Population Survey public data series.

<strong>Table 5.3. Unemployment rate, by education and race and ethnicity, 2000–2011.</strong>Underlying data are basic monthly Current Population Survey microdata.

<strong>Table 5.4. Unemployment rate, by gender and education, 2000–2011.</strong> Underlying data are basic monthly Current Population Survey microdata.

<strong>Table 5.5.</strong> <strong>Decline in the labor force participation rate from 1989 to 2011 and its possible effect on the unemployment rate in 2011, by gender and age.</strong>Underlying data are basic monthly Current Population Survey microdata. The counterfactual 2011 labor force participation rate is what the labor force participation rate would have been in 2011 if the labor force participation rate of each of 30 gender/age/education cells had continued on the same linear trend from 2007 to 2011 that they followed from 1989 to 2007, but if the relative sizes of those cells evolved as they actually did. (Note, there are three age groups: 16–24, 25–54, and 55+; and five education groups: less than high school, high school, some college, college degree, and advanced degree. The table presents aggregated results by gender and age.) The counterfactual 2011 unemployment rate is what the unemployment rate in 2011 would have been if the workers making up the difference between the actual and the counterfactual 2011 labor force participation rate were in the labor force and unemployed instead of out of the labor force.

<strong>Table 5.6. Underemployment, 2000–2011.</strong> Underlying data are from the Current Population Survey public data series. <em>Involuntary part time</em> refers to those who work part time for economic reasons, i.e., those who want and are available for full-time work, but who have had to settle for a part-time schedule. <em>Marginally attached </em>refers to those who are currently neither working nor looking for work but indicate that they want and are available for a job and have looked for work sometime in the past year.

<strong>Table 5.7.</strong> <strong>Long-term unemployment, by demographic group, education, and occupation, 2000–2011.</strong>Underlying data are from the Bureau of Labor Statistics Current Employment Statistics public data series.

<strong>Table 5.8. Industry distribution and job loss, by gender, 2007–2011.</strong>Underlying data are from the Bureau of Labor Statistics Current Employment Statistics public data series.

<strong>Figure 5A.</strong> <strong>Jobs needed each month to hold steady and actual monthly job growth, 1969–2011.</strong> <em>Actual monthly job growth,</em> the number of jobs added per month on average, comes from the Bureau of Labor Statistics Current Employment Statistics (CES) public data series. <em>Jobs needed each month to hold steady </em>is the number of jobs needed per month on average in a given year to maintain the same ratio of payroll jobs to the working-age population that prevailed at the end of the prior year (payroll jobs data come from the CES, and the size of the working-age population age 16 and older comes from the Current Population Survey public data series). A three-year rolling average of the working-age population in December is used because of large year-to-year variability in the population growth rate as measured by the CPS.

<strong>Figure 5B. Distribution of employment, by industry, selected years, 1979–2011 (and 2020 projections).</strong> Underlying data for 1979–2011 are from the Bureau of Labor Statistics Current Employment Statistics public data series. Underlying data for 2020 are from the Employment Projections program, Table 2.1, “Employment by Major Industry Sector.”

<strong>Figure 5C.</strong> <strong>Distribution of employment, by firm size, 2011Q1.</strong> Underlying data are from the Bureau of Labor Statistics Business Employment Dynamicsprogram, <em>National Firm Size Data—Supplemental Firm Size Class Tables</em>, Table F, “Distribution of Private Sector Employment by Firm Size Class, Not Seasonally Adjusted.”

<strong>Figure 5D.</strong> <strong>Job gains, losses, and net employment change, by firm size, 2000–2011.</strong> Underlying data are from the Bureau of Labor Statistics Business Employment Dynamics program, <em>National Firm Size Data—Size Class 1 Tables</em>, Table 1, “Private Sector Firm-level Gross Job Gains and Job Losses: Seasonally Adjusted, Dynamic Method.”

<strong>Figure 5E.</strong> <strong>Distribution of employment, by occupation, selected years, 1989–2011. </strong>Underlying data are from the Current Population Survey public data series, Historical Table A-13, “Employed and Unemployed Persons by Occupation, Not Seasonally Adjusted.” Service occupations include health care support, protective service, food preparation and serving, building and grounds cleaning and maintenance, and personal care and service occupations.

<strong>Figure 5F. Good jobs as a share of total employment, all workers and by gender, and output per worker, selected years, 1979–2010.</strong>Good jobs shares are from Schmitt and Jones (2012), and output per worker is from the Bureau of Labor Statistics Labor Productivity and Costs program (unpublished Total Economy Productivity data provided to the authors upon request). Good jobs are defined as those that pay at least $18.50 per hour (the median male hourly wage in 1979 adjusted to 2010 dollars), have employer-provided health insurance where the employer pays at least some of the premium, and an employer-sponsored pension plan, including 401(k) and similar defined-contribution plans.

<strong>Figure 5G. Unemployment rate, 1948–2011.</strong>Underlying unemployment data are from the Current Population Survey public data series.

<strong>Figure 5H.</strong> <strong>Unemployment rate (actual and holding age distribution constant), 1979–2011.</strong> Underlying data are from the Current Population Survey public data series. The unemployment rate holding the age distribution constant since 1979 is the result of a simple exercise showing what the unemployment rate would be if the distribution of the labor force across age categories (ages 16–24, 25–34, 35–44, 45–54, and 55 and older) had not changed since January 1979, but the unemployment rates within each age category evolved as they actually did from January 1979 to December 2011.

<strong>Figure 5I.</strong> <strong>Unemployment rate, by race and ethnicity, 1979–2011.</strong> Underlying data are basic monthly Current Population Survey microdata. As with other CPS microdata analyses presented in the book, race/ethnicity categories are mutually exclusive (i.e., white non-Hispanic, black non-Hispanic, and Hispanic any race).

<strong>Figure 5J. Unemployment rates of foreign-born and native-born workers, 1994–2011.</strong> Underlying data are basic monthly Current Population Survey microdata.

<strong>Figure 5K.</strong> <strong>Share of unemployed people with unemployment insurance benefits, 1989–2011.</strong> Underlying data are from the Current Population Survey public data series and the U.S. Department of Labor’s Unemployment Insurance Program Statistics, ”Persons Claiming UI Benefits in Federal Programs (Expanded)” [Excel spreadsheet]. Extended benefits refer to those extended by Congress during downturns beyond the regular state-financed benefits. Shares are calculated by dividing the number of persons claiming regular benefits by the total number of unemployed persons, and by dividing the total number of persons claiming extended benefits or regular benefits by the total number of unemployed persons. Weekly unemployment insurance claims data are converted into monthly data from January 1989 to December 2011.

<strong>Figure 5L. Labor force participation rate, by age and gender, 1959–2011.</strong> Underlying data are from the Current Population Survey public data series.

<strong>Figure 5M. Employment-to-population ratio, age 25–54, by gender, 1989–2011. </strong>Underlying data are from the Current Population Survey public data series.

<strong>Figure 5N. Underemployment rate, by race and ethnicity, 2000–2011.</strong> Underlying data are basic monthly Current Population Survey microdata. As with other CPS microdata analyses presented in the book, race/ethnicity categories are mutually exclusive (i.e., white non-Hispanic, black non-Hispanic, and Hispanic any race).

<strong>Figure 5O. Long-term unemployment, 1948–2011.</strong>Underlying data are from the Current Population Survey public data series.

<strong>Figure 5P. Unemployment rate, average monthly and over-the-year, 2000–2010.</strong>Average monthly unemployment rate underlying data are from the Current Population Survey public data series, and over-the-year unemployment underlying data are from the U.S. Bureau of Labor Statistics <em>Work Experience of the Population</em> (annual economic news release).

<strong>Figure 5Q. Job-seekers ratio, Dec. 2000–Dec. 2011.</strong>Job openings data are from the U.S. Bureau of Labor Statistics Job Openings and Labor Turnover Survey, and unemployment data are from the Current Population Survey public data series.

<strong>Figure 5R. Voluntary quits, Dec. 2000–Dec. 2011.</strong>Underlying data are from the Bureau of Labor Statistics Job Openings and Labor Turnover Survey.

<strong>Figure 5S.</strong> <strong>Job change since the start of each of the last four recessions.</strong> Underlying data are from the Bureau of Labor Statistics Current Employment Statistics public data series. Data for each recession are indexed by the number of jobs in the first month of the recession. Monthly data span July 1989–December 2011.

<strong>Figure 5T.</strong> <strong>Job change since the start of each of the last four recoveries (all, private sector, and public sector).</strong> Underlying data are from the Bureau of Labor Statistics Current Employment Statistics public data series. Data for each recession are indexed by the number of jobs in the first month of the recession’s recovery. Monthly data span July 1989–December 2011.

<strong>Figure 5U. Job change, by gender, in the Great Recession and its aftermath (Dec. 2007–Dec. 2011).</strong>Underlying data are from the Bureau of Labor Statistics Current Employment Statistics public data series. Data for each gender are indexed by the number of jobs held by workers of that gender in the first month of the recession.

<strong>Figure 5V.</strong> <strong>Simulated job change by gender in the Great Recession and its aftermath (Dec. 2007–Dec. 2011), controlling for industry.</strong>Underlying data are from the Bureau of Labor Statistics Current Employment Statistics public data series. The graph presents the results of an exercise showing how employment of men and women would have changed over the four-year period if, in December 2007, men and women had had the same industry distribution but if job changes by gender within each industry had evolved as they actually did between December 2007 and December 2011.

<strong>Figure 5W. Unemployed workers and job openings, by industry, 2011.</strong>Underlying data are from the Bureau of Labor Statistics Job Openings and Labor Turnover Survey and the Current Population Survey public data series.

<strong>Figure 5X. Unemployed workers, by occupation, 2007 and 2011.</strong>Underlying data are from basic monthly Current Population Survey microdata.

<strong>Figure 5Y. Unemployment rate, by education, 2007 and 2011.</strong>Underlying data are basic monthly Current Population Survey microdata.

<strong>Figure 5Z. Labor force status of involuntarily displaced workers, 1984–2010.</strong>Underlying data are from Farber (2011), Table 6, “Post-displacement Labor Force Status, 1984–2010.”

<strong>Figure 5AA. Average decline in weekly earnings for involuntarily displaced full-time workers who found new work, 1984–2010. </strong>Underlying data are from Farber (2011), Table 16, “Proportional Change in Real Weekly Earnings, Full-Time Job Losers.”

<strong>Table 6.1. Distribution of income compared with distribution of wealth, 2010.</strong> The table is based on unpublished analysis of 2010 Survey of Consumer Finances (SCF) data prepared in 2012 by Edward Wolff for the Economic Policy Institute. The definition of wealth used in this analysis of the SCF is the same definition of wealth used in the analysis of the SCF conducted by Bricker et al. (2012), except that the Bricker et al. analysis includes vehicle wealth, while this analysis does not.

<strong>Table 6.2. Change in wealth groups’ shares of total wealth, 1962–2010.</strong> See note to Table 6.1.

<strong>Table 6.3. Change in average wealth, by wealth group, 1962–2010.</strong> See note to Table 6.1.

<strong>Table 6.4. Share of households with low net worth, 1962–2010.</strong> See note to Table 6.1.

<strong>Table 6.5. Median household wealth, and share of households with zero or negative wealth, by race and ethnicity, 1983–2010.</strong> See note to Table 6.1.

<strong>Table 6.6. Wealth groups’ shares of assets, by asset type, 2010.</strong> See note to Table 6.1.

<strong>Table 6.7. Average household assets, by wealth group and asset type, 1962–2010.</strong> See note to Table 6.1.

<strong>Table 6.8. Average and median household assets, by race/ethnicity and asset type, 1983–2010.</strong> See note to Table 6.1.

<strong>Table 6.9. Share of households owning stock, 1989–2010.</strong> See note to Table 6.1.

<strong>Table 6.10. Average household debt, assets, and net worth, by wealth group, 1962–2010.</strong> See note to Table 6.1.

<strong>Table 6.11. Median household debt, by race and ethnicity, 1983–2010.</strong> See note to Table 6.1.

<strong>Table 6.12. Distribution of debt by its purpose, 1989–2010.</strong> Data for years 2001–2010 are from Bricker et al. (2012). Data for prior years are from the Federal Reserve Board’s Survey of Consumer Finances, <em>Tables Based on the Internal Data.</em>

<strong>Table 6.13. Household financial obligations as a share of disposable personal income, for renters and homeowners, 1980–2011. </strong>Data refer to annual averages from the Federal Reserve Board (FRB), “Household Debt Service and Financial Obligations Ratios.” Per the FRB, the <em>financial obligations ratio (FOR)</em> adds automobile lease payments, rental payments on tenant-occupied property, homeowners’ insurance, and property tax payments to the debt service ratio (an estimate of the ratio of debt payments on outstanding mortgage and consumer debt, to disposable personal income). The <em>homeowner mortgage FOR</em> includes payments on mortgage debt, homeowners’ insurance, and property taxes, while the <em>homeowner consumer FOR</em> includes payments on consumer debt and automobile leases.

<strong>Table 6.14. Household debt service as a share of family income, by income group, 1989–2010.</strong> Data are from Bricker et al. (2012), Table 17.

<strong>Table 6.15. Share of households with high debt burdens, by income group, 1989–2010.</strong> Data are from Bricker et al. (2012), Table 17.

<strong>Table 6.16. Share of households late paying bills, by income group, 1989–2010.</strong> Data are from Bricker et al. (2012), Table 17.

<strong>Table 6.17. Median and mean wealth per adult in 20 advanced countries, 2011.</strong> Data are from the Credit Suisse Research Institute’s <em>Global Wealth Databook 2011</em>. Note that in international comparisons of income it is standard practice (including at EPI) to convert currencies using “purchasing power parity” (PPP) exchange rates instead of market exchange rates. PPPs are based on the price of buying a given “basket” of goods and services in each country, thereby equalizing the purchasing power of currencies. It should be noted that for these data, market exchange rates, not PPP exchange rates, are used to convert currencies to U.S. dollars. The authors of the report argue that there is a case to be made for using market exchange rates for international comparisons of wealth because in every country a large share of personal wealth is owned by households in the top few percentiles of the distribution, and these households tend to move their assets across borders with relative frequency. Results are not available using PPP exchange rates. It also should be noted that the ratio of mean to median is the same regardless of what exchange rates are used.

<strong>Figure 6A. Average household net worth, net financial assets, and net tangible assets, 1965–2012.</strong> Data for net worth and assets are from the Federal Reserve Board’s Flow of Funds data, Table B.100, “Balance Sheet of Households and Nonprofit Organizations.” The data were adjusted for inflation using the CPI-U-RS (Consumer Price Index Research Series Using Current Methods), and divided by the number of U.S. households based on Census Bureau data. The household data are from the Current Population Survey/Housing Vacancy Survey Historical Tables, Table 7, “Annual Estimates of the Housing Inventory: 1965 to Present” (http://www.census.gov/hhes/www/housing/hvs/historic/index.html). The number of “owner occupied” homes was taken as a percentage of “total occupied” homes to calculate a percentage of homeownership.

<strong>Figure 6B. Share of total household wealth growth accruing to various wealth groups, 1983–2010.</strong> Data are derived from Table 6.3.

<strong>Figure 6C. Ratio of average top 1% household wealth to median wealth, 1962–2010.</strong> Data are derived from Table 6.3.

<strong>Figure 6D. Average annual net worth of “Forbes 400” wealthiest individuals, 1982–2011. </strong>Data for 1982 to 1999 are adapted from Broom and Shay (2000) Table 2, “‘Forbes 400’ Individual Fortunes.” Data from 2000 to 2011 are from Forbes annual lists of the richest 400 Americans. All data are adjusted to 2011 dollars using the CPI-U-RS.

<strong>Figure 6E. Median household wealth by race and ethnicity, 1983–2010.</strong> See note to Table 6.5.

<strong>Figure 6F. U.S. stock market, 1955–2011. </strong>Data on the Standard &amp; Poor’s composite index of the 500 largest U.S. firms (the S&amp;P 500) are from the <em>Economic Report of the President</em> (Council of Economic Advisers 2012), tables B-95, “Historical Stock Prices and Yields, 1949–2003,” and B-96, “Common Stock Prices and Yields, 2000–2011,” deflated by the CPI-U-RS in 2011 dollars and indexed to 1960=100.

<strong>Figure 6G. Wealth groups’ shares of total household stock wealth, 1983–2010.</strong> Data are derived from Table 6.6; see table note to Table 6.6.

<strong>Figure 6H. Annual homeownership rate, 1965-2011.</strong> Annual data are from the Current Population Survey/Housing Vacancy Survey, <em>Historical Tables</em>, Table 7, “Annual Estimates of the Housing Inventory: 1965 to Present,” http://www.census.gov/hhes/www/housing/hvs/historic/index.html. To calculate the rate of homeownership, the number of owner-occupied homes was taken as a percentage of total occupied homes.

<strong>Figure 6I. Homeownership rate, by income group, 2009.</strong> Data are from the Census Bureau’s American Housing Survey, <em>National Tables</em>, Table 3-12, “Owner Occupied Units,” http://www.census.gov/housing/ahs/data/ahs2009.html, most recently published in 2009. Due to budget constraints, the two-year schedule for this survey was delayed in 2012, and 2011 data were not available in time for this publication.

<strong>Figure 6J. Homeownership rate, by race and ethnicity, 1975–2011.</strong> Data prior to 1994 are taken from the Current Population Survey (CPS) Annual Social and Economic Supplement, provided by the Census Bureau upon request. Data from 1994 onward are taken from the CPS/Housing Vacancy Survey, <em>Annual Statistics: 2011</em>, Table 22, “Homeownership Rates by Race and Ethnicity of Householder” (http://www.census.gov/hhes/www/housing/hvs/annual11/ann11ind.html). As with other CPS microdata analyses presented in this book, race/ethnicity categories are mutually exclusive (i.e., white non-Hispanic, black non-Hispanic, and Hispanic any race).

<strong>Figure 6K. Home prices, 1953–2012.</strong> Home price data are from Robert Shiller, of Yale University, who publishes a quarterly series of home price data, which was featured in his book <em>Irrational Exuberance </em>(http://www.econ.yale.edu/~shiller/data.htm). The home price index is set to 1997Q1=100.

<strong>Figure 6L. Total homeowner equity as a share of total home values, 1969–2011.</strong> Data are from the Federal Reserve Board’s Flow of Funds data, Table B.100, “Balance Sheet of Households and Nonprofit Organizations.”

<strong>Figure 6M. Foreclosures per 1,000 owner-occupied dwellings, 2000–2011.</strong> Data on foreclosures are from the Federal Reserve Bank of New York’s <em>Quarterly Report on Household Debt and Credit</em>; data series “Number of Consumers with New Foreclosures” (http://www.newyorkfed.org/newsevents/news/research/2012/an120227.html). The number of owner-occupied dwellings was taken from the Current Population Survey/Housing Vacancy Survey, <em>Historical</em> <em>Tables</em>, Table 8, “Quarterly Estimates of the Housing Inventory: 1965 to Present” (http://www.census.gov/hhes/www/housing/hvs/historic/index.html).

<strong>Figure 6N. Enrollment in defined-benefit versus defined-contribution pension plans among workers with pension coverage, 1983 and 2010.</strong> Figure produced from data in Munnell (2012), Figure 4.

<strong>Figure 6O. Debt as a share of disposable personal income, all and by type of debt, 1946–2011.</strong> Data on disposable personal income, consumer credit liability, total liabilities, and mortgage liabilities are from the Federal Reserve Board’s Flow of Funds data, Table B.100, “Balance Sheet of Households and Nonprofit Organizations.” Data on home equity loans are from Flow of Funds data, Table L.218, “Home Mortgages,” and are unavailable prior to 1990. The various liabilities are taken as shares of disposable personal income for display in the graphs.

<strong>Figure 6P. Consumer bankruptcies per 1,000 adults, 1989–2011.</strong> Data on bankruptcies are American Bankruptcy Institute Annual and Quarterly U.S. Bankruptcy Statistics, “Annual Business and Non-business Filings by Year” (http://www.abiworld.org/Content/NavigationMenu/NewsRoom/BankruptcyStatistics/Bankruptcy_Filings_1.htm). Data on the adult population are calculated with labor force statistics from the Current Population Survey, Civilian Non-institutional Population Series, Ages 18 and Over.

<strong>Figure 6Q. Median wealth per adult in 20 advanced countries, 2011.</strong> See note to Table 6.17.

<strong>Table 7.1. Comparison of poverty measures. </strong>Table is adapted from Short (2011), “Resource Estimates” table.

<strong>Table 7.2. Contribution of hours versus hourly wages to annual wage growth for working-age households, selected years, 1979–2007. </strong>See note to Table 2.16.

<strong>Table 7.3. Impact of changes in U.S. economic and demographic composition on the poverty rate, selected periods, 1979–2010.</strong> Underlying data are from Current Population Survey Annual Social and Economic Supplement microdata; see Appendix A for details. The methodology for this decomposition is taken from Danziger and Gottschalk (1995, Chapter 5), which explores the role of changes in socioeconomic characteristics (e.g., changes in average income, changes in income inequality, and demographic changes such as the change in racial groups’ shares of the overall population) on the poverty rate (using the official poverty rate) between any two years. We focus specifically on the 1979–1989, 1989–2000, 2000–2007, 2007–2010, and 1979–2007 periods. To examine the impact of average income of the U.S. population on the poverty rate, we assign the average real income growth across the period to be the growth for all individuals between years t0 and t1 and simulate a new poverty rate. This procedure holds the shape of the distribution (inequality) constant in t0 while allowing incomes to grow equally for all individuals. This simulated poverty rate for t1 is then compared to the actual poverty rate in year t0, and the percentage-point difference is the change in the mean, i.e. the impact of income growth. The change due to income inequality is the percentage-point difference between the simulated poverty distribution in t1 and the actual poverty rate in t1.

We repeat this exercise using the demographic composition of each variable of interest to see the effect of these demographic changes on the overall poverty rate. First we calculate the weight of each demographic factor (such as individuals with college degrees) by its population share and simulate the poverty rate in t1 for all persons between t0 and t1, allowing for income to grow equally among all families and holding the demographic composition of the population in t0 constant. Then, we calculate a second simulated rate that incorporates both the mean income growth and the demographic changes across the period. The difference between these two simulated rates in t1 is the percentage point-change in the poverty rate due to demographic changes.

The interaction, or error, term states to what degree the demographic variables are conflated, which could lead to bias in measurement of a factor’s impact. Since our error term is negative and relatively small (-.4 from 1979–2007), the reported relationship might slightly overstate the degree to which the simulated income decreases the poverty rate for each demographic group, but it is not enough to change the story.

<strong>Figure 7A. Poverty and twice-poverty rates, 1959–2010. </strong>Underlying data are from Current Population Survey Annual Social and Economic Supplement <em>Historical Poverty Tables</em>, Table 2, “Poverty Status, by Family Relationship, Race, and Hispanic Origin,” and Table 5, “Percent of People by Ratio of Income to Poverty Level.”

<strong>Figure 7B. Poverty rate, by age, 1959–2010. </strong>Underlying data are from Current Population Survey Annual Social and Economic Supplement <em>Historical Poverty Tables</em>, Table 3, “Poverty Status, by Age, Race, and Hispanic Origin” and Current Population Survey Annual Social and Economic Supplement microdata (see Appendix A for details).

<strong>Figure 7C. Poverty rate, by race and ethnicity, nativity, and citizenship status, 1973–2010. </strong>Underlying data are from Current Population Survey Annual Social and Economic Supplement <em>Historical Poverty Tables</em>, Table 2, “Poverty Status, by Family Relationship, Race, and Hispanic Origin” and Table 23, “People in Poverty by Nativity.” As with most other CPS data analyses presented in the book, race/ethnicity categories are mutually exclusive (i.e., white non-Hispanic, black non-Hispanic, and Hispanic any race).

<strong>Figure 7D. Poverty rate, by race and ethnicity and age, 2010. </strong>Underlying data are from Current Population Survey Annual Social and Economic Supplement (CPS-ASEC) <em>Historical Poverty Tables</em>, Table 3, “Poverty Status, by Age, Race, and Hispanic Origin” and from CPS-ASEC microdata; see Appendix A for details. As with most other CPS data analyses presented in the book, race/ethnicity categories are mutually exclusive (i.e. white non-Hispanic, black non-Hispanic, and Hispanic any race).

<strong>Figure 7E. Poverty rates of various types of families, 1959–2010. </strong>Underlying data are from Current Population Survey Annual Social and Economic Supplement <em>Historical Poverty Tables</em>, Table 4, “Poverty Status, by Type of Family, Presence of Related Children, Race and Hispanic Origin.”

<strong>Figure 7F. Length of time in poverty over a two-year period, 2008–2009. </strong>Underlying data are from Survey of Income and Program Participation microdata (2008 panel).

<strong>Figure 7G. Share of the poor in “deep poverty,” 1975–2010. </strong>Underlying data are from Current Population Survey Annual Social and Economic Supplement <em>Historical Poverty Tables</em>, Table 2, “Poverty Status, by Family Relationship, Race, and Hispanic Origin,” and Table 22, “Number of People Below 50 Percent of Poverty Level.”

<strong>Figure 7H. Poverty rate, official and under the Supplemental Poverty Measure, by age group, 2010. </strong>Underlying data are from the U.S. Census Bureau’s <em>Current Population Reports </em>(Short 2011), Table 1, “Number and Percent of People in Poverty by Different Poverty Measures: 2010.”<strong> </strong>

<strong>Figure 7I. Official and relative poverty rate, 1979–2010. </strong>Underlying data are from Current Population Survey Annual Social and Economic Supplement <em>Historical Poverty Tables</em>, Table 2, “Poverty Status, by Family Relationship, Race, and Hispanic Origin,” and Current Population Survey Annual Social and Economic Supplement microdata; see Appendix A for details. To be consistent with international comparisons, median income includes noncash transfers such as food stamps and housing subsidies.

<strong>Figure 7J. Demographic characteristics of poverty-level-wage workers vs. non-poverty-level-wage workers, 2011. </strong>Underlying data are from Current Population Survey Outgoing Rotation Groups microdata; see Appendix B for details. As with most other CPS microdata analyses presented in the book, race/ethnicity categories are mutually exclusive (i.e. white non-Hispanic, black non-Hispanic, and Hispanic any race).

<strong>Figure 7K. Industry, occupation, and union status of poverty-level-wage workers vs. non-poverty-level-wage workers, 2011. </strong>Underlying data are from Current Population Survey Outgoing Rotation Groups microdata; see Appendix B for details. Occupations do not sum to 100 percent because the figure excludes the “Other Occupations” category, which constitutes less than 2 percent of the workforce.

<strong>Figure 7L. Share of poverty-level-wage and non-poverty-level-wage workers with employer-sponsored health insurance and pension coverage, 2010. </strong>Underlying data are from Current Population Survey Annual Social and Economic Supplement microdata; see Appendix A for details. The analysis includes workers in both the private and public sectors and does not have age limits or work requirements. Coverage is defined as being included in an employer-sponsored plan for which the employer paid for at least some of the coverage.

<strong>Figure 7M. Poverty rate, actual and simulated, 1959–2010. </strong>Underlying data are from Current Population Survey Annual Social and Economic Supplement <em>Historical Poverty Tables</em>, Table 2, “Poverty Status, by Family Relationship, Race, and Hispanic Origin,” and Table 4, “Poverty Status, by Type of Family, Presence of Related Children, Race and Hispanic Origin,” and from Bureau of Economic Analysis National Income Product Accounts public data, Table 7.1, “Selected Per Capita Product and Income Series in Current and Chained Dollars.” The analysis is an adaptation of analysis by Danziger and Gottschalk (1995), whose method was to regress the poverty rate of the growth of real per capita gross domestic product from 1959–1973 and then simulate poverty rates based on that simple model. The link between GDP and poverty in the earlier period (1959–1973) and the potential for GDP to eradicate poverty by the 1980s holds true for alternative specifications including using only the under-age-65 poverty rate (to remove elderly, the main recipients of Social Security, also growing over this period) and controlling for one target demographic: female headed families.

Figure 7N. Change in productivity, 20th-percentile wages, unemployment, and poverty, selected periods, 1979–2010. Productivity data, which measure output per hour, are from the Bureau of Labor Statistics Major Sector Productivity and Costs Index public data series; the figure shows the average annual growth rate of productivity over the periods covered. The figure also shows the average annual growth rate of wages at the 20th percentile of the wage distribution for the given periods, using data from Current Population Survey Outgoing Rotations Group microdata; see Appendix B for details. The percentage-point changes in the unemployment rate across the periods shown come from the monthly Current Population Survey public data series, while percentage-point changes in the poverty rate come from Current Population Survey Annual Social and Economic Supplement <em>Historical Poverty Tables</em>, Table 2, “Poverty Status, by Family Relationship, Race, and Hispanic Origin.”

<strong>Figure 7O. Increase in wages from a 1-percentage-point decline in the unemployment rate, by gender. </strong>Estimates use Current Population Survey Outgoing Rotation Group microdata (see Appendix B), and are computed based on a model employed by Katz and Krueger (1999). Annual changes in log wages are regressed on unemployment, lagged log-changes in the CPI-U-RS (but, following Katz and Krueger the coefficient on this is constrained to equal 1), lagged productivity growth, and dummies for 1989–1995, 1996–2000, and 2001–2007 (excluded period is 1979–1988). The sample covers the years 1979–2007.

<strong>Figure 7P. Impact of changes in family structure on the poverty rate, selected periods, 1979–2010. </strong>The figure looks at the overall composition of family structure in the United States (e.g., the share of families headed by a single mother) and measures how much the change in the composition has affected the poverty rate in given periods. For more information on the methodology underlying the figure, see the note to Table 7.3.

<strong>Figure 7Q. Impact of changes in U.S. economic and demographic composition on the poverty rate, 1979–2007. </strong>See note to Table 7.3.

<strong>Figure 7R. Per capita Social Security expenditures and the elderly poverty rate, 1959–2010. </strong>Underlying data are from Current Population Survey Annual Social and Economic Supplement <em>Historical Poverty Tables</em>, Table 2, “Poverty Status, by Family Relationship, Race, and Hispanic Origin,” and Table 3, “Poverty Status, by Age, Race, and Hispanic Origin.” Data are also from the Social Security Administration <em>Trustees Report 2009: Annual Statistical Supplement</em>, Table 4a, “Old-Age and Survivors Insurance Trust Fund Expenditures”

<strong>Figure 7S. Poverty rate absent targeted government programs, by age group, 2010. </strong>Underlying data are from Short (2011), Table 3a, “Effect of Excluding Individual Elements on SPM Rates: 2010.”

<strong>Figure 7T.</strong> <strong>Share of bottom-fifth household income accounted for by wages, cash transfers, and in-kind income, 1979–2007.</strong> Underlying data are from the 2010 Congressional Budget Office, <em>Average Federal Taxes by Income Group</em>, “Sources of Income for all Households, by Household Income Category, 1979–2007” [Excel spreadsheet]. The Congressional Budget Office definition of in-kind income includes employer-paid health insurance premiums, food stamps, school lunches and breakfasts, housing assistance, energy assistance, and the fungible value of Medicare and Medicaid, as estimated by the Current Population Survey. CBO’s definition of cash transfers includes payments from Social Security, unemployment insurance, Supplemental Security Income, Aid to Families with Dependent Children, Temporary Assistance for Needy Families, veterans’ benefits, and workers’ compensation.

<strong>Figure 7U. Earnings at the 10th percentile as a share of median worker earnings in selected OECD countries, late 2000s. </strong>Underlying data are metadata from the Organisation for Economic Co-operation and Development’s <em>Distribution of Gross Earnings of Full-time Employees and Gender Wage Gap </em>database. Earnings for all countries are defined as gross earnings for full-time, full-year workers, with the exception of Denmark, which is for all workers, the Netherlands, which is for full time, full-year equivalent workers, and Switzerland, which is net earnings for full-time workers. The shares are earnings at the 10th percentile as a share of the median earnings in each country’s respective currency.

<strong>Figure 7V. Earnings at the 10th percentile in selected OECD countries relative to the United States, late 2000s. </strong>Underlying data are metadata from the Organisation of Economic Co-operation and Development’s <em>Distribution of Gross Earnings of Full-time Employees and Gender Wage Gap </em>database. See note for Figure 7U on definition of earnings. Data for earnings at the 10th percentile are converted into weekly earnings and are then converted into equivalent U.S. dollars using a purchasing power parity index from the International Monetary Fund <em>World Economic Outlook</em> <em>Database</em>. The figure shows the share of each country’s 10th percentile earnings relative to the 10th percentile earnings in the United States.

<strong>Figure 7W. Relative poverty rate in the United States and selected OECD countries, late 2000s. </strong>Underlying data are from the Organisation for Economic Co-operation and Development’s <em>Stat Extracts</em> public data series. Household-size-adjusted income, or equivalent income, is household income divided by the square root of the household size. Countries were chosen based on their productivity per worker hour using the “PPP Converted GDP Laspeyres Per Hour Worked by Employees at 2005 Constant Prices” series from <em>Penn World Table Version 7.0 </em>(Heston, Summers, and Aten 2011). We chose to exclude countries whose productivity is less than half that of the United States. The OECD data base uses slightly different methods than that found in 7I (e.g., its handling of taxes and transfers are different), therefore, the relative rates for the United States are not exactly the same.

<strong>Figure 7X. Child poverty rate in selected developed countries, 2009. </strong>Underlying data are from UNICEF Innocenti Research Centre Report Card 10 (Adamson 2012)<em>,</em> Figure 1b, “Child Poverty Rate.” The poverty rate is the percentage of children (age 0–17) living in households with equivalent income lower than 50 percent of the national median, where equivalent income is disposable income, adjusted for family size and composition. UNICEF uses a modified equivalence scale to adjust for household size by weighting the first adult in the household by 1, the subsequent adults by .5, and children under age 14 by .3, then summing the weights up and dividing total household income by the total weight. We chose countries based on their productivity per worker hour using the “PPP Converted GDP Laspeyres Per Hour Worked by Employees at 2005 Constant Prices” series from <em>Penn World Table Version 7.0 </em>(Heston, Summers, and Aten 2011) and excluded countries whose productivity is less than half that of the United States.

<strong>Figure 7Y. Child poverty gap in selected developed countries, 2009. </strong>Underlying data are from UNICEF Innocenti Research Centre Report Card 10 (Adamson 2012), Figure 7, “The Poverty Gap.” The child poverty gap is the distance between the poverty line and the median family income of children below the poverty line, expressed as a percentage of the poverty line. This is calculated by lining up all individuals in households by household-size-adjusted income (with children taking their family income value) and then locating the poverty line, which is 50 percent of national median income. UNICEF uses a modified equivalence scale to adjust for household size by weighting the first adult in the household by 1, the subsequent adults by .5, and children under age 14 by .3, then summing the weights up and dividing total household income by the total weight. The median income of children below the poverty line is then calculated. Then the gap between the poverty line and the median income of children is then taken as a share of the poverty line. For example, for a country with a median income of $50,000, the poverty line is $25,000. If the median income for children living below $25,000 is $15,000, the difference is $25,000<strong>-</strong>$15,000 = $10,000. This difference, taken as a share of the poverty line, yields a child poverty gap of $10,000/$25,000 (40 percent). We chose countries from the UNICEF list based on their productivity per worker hour using the “PPP Converted GDP Laspeyres Per Hour Worked by Employees at 2005 Constant Prices” series from <em>Penn World Table Version 7.0 </em>(Heston, Summers, and Aten 2011), and excluded countries whose productivity is less than half that of the United States.

<strong>Figure 7Z. Extent to which taxes and transfer programs reduce the relative poverty rate, selected developed OECD countries, late 2000s.</strong> Underlying data are from the Organisation for Economic Co-operation and Development’s <em>Stat Extracts </em>public data series. Household-size-adjusted income, or equivalent income, is household income divided by the square root of the household size. We chose countries based on their productivity per worker hour using the “PPP Converted GDP Laspeyres Per Hour Worked by Employees at 2005 Constant Prices” series from <em>Penn World Table Version 7.0 </em>(Heston, Summers, and Aten 2011), and excluded countries whose productivity is less than half that of the United States.

<strong>Figure 7AA. Social expenditure and relative poverty rates selected in OECD countries, late 2000s.</strong> Underlying data are from the Organisation for Economic Co-operation and Development’s<em> Stat Extracts </em>public data series. The relative poverty rate is the share of individuals living in households with income below half of household-size-adjusted median income, which is household income divided by the square root of the household size. We chose countries based on their productivity per worker hour using the “PPP Converted GDP Laspeyres Per Hour Worked by Employees at 2005 Constant Prices” series from the <em>Penn World Table Version 7.0</em> (Heston, Summers, and Aten 2011), and excluded countries whose productivity is less than half that of the United States.