Shared prosperity

Lane Kenworthy, The Good Society
July 2016

As a society gets richer, its citizens’ living standards should rise. The poorest needn’t benefit the most; equal rates of improvement may be good enough. We might not even mind if the wealthiest benefit a bit more than others; a little increase in income inequality is hardly catastrophic. But in a good society, those in the middle and at the bottom ought to benefit significantly from economic growth. When the country prospers, everyone should prosper.1

How has the United States fared in recent decades? How has it done compared to other affluent countries? I examine the least well-off in “A Decent and Rising Income Floor.” Here I focus on households in the broad middle.


The United States is among the richest of the world’s longstanding democracies. Its GDP per capita trails only Norway’s.2 How effectively is that prosperity shared? Figure 1 shows disposable household income at the 75th percentile. (About 75% of households have an income that’s lower, and 25% have an income that’s higher.) At this point on the ladder, US households are richer than their counterparts in any other affluent nation. And since the 1970s their income has increased by about $15,000 — a modest improvement, but hardly a trivial one.

Figure 1. Seventy-fifth-percentile household income
Posttransfer-posttax household income. The incomes are adjusted for household size and then rescaled to reflect a three-person household, adjusted for inflation, and converted to US dollars using purchasing power parities. “k” = thousand. The lines are loess curves. Data sources: Luxembourg Income Study; OECD.

Figure 2 shows income at the median — the 50th percentile.3 (Figure A1 in the appendix has a separate graph for each country.) Here America is near the top in the cross-country ranking. But our median income is below that of Norway and Switzerland, and it has increased only a little since the 1970s, so other nations have been catching up.4

Figure 2. Median household income
Posttransfer-posttax household income. The incomes are adjusted for household size and then rescaled to reflect a three-person household, adjusted for inflation, and converted to US dollars using purchasing power parities. “k” = thousand. The lines are loess curves. Data sources: Luxembourg Income Study; OECD.

Figure 3 shows income at the 25th percentile. At this point on the ladder the United States is in the middle of the pack among the world’s affluent countries. And we’ve had little increase in recent decades. Quite a few other nations have caught up with us or pulled ahead.

Figure 3. Twenty-fifth-percentile household income
Posttransfer-posttax household income. The incomes are adjusted for household size and then rescaled to reflect a three-person household, adjusted for inflation, and converted to US dollars using purchasing power parities. “k” = thousand. The lines are loess curves. Data sources: Luxembourg Income Study; OECD.


In the period between World War II and the mid-to-late 1970s, economic growth was good for Americans in the middle. Figure 4 shows that as GDP per capita increased, so did family income at the 80th percentile, the 50th percentile (the median), and the 20th percentile.5 Indeed, they moved virtually in lockstep. Since then, however, household income has become decoupled from economic growth. As the economy has grown, relatively little of that growth has reached households in the middle, particularly those in the lower-middle.

Figure 4. GDP per capita and middle-class family income
P20 is the 20th percentile on the income ladder; P50 is the 50th percentile (median); P80 is the 80th percentile. Each series is displayed as an index set to equal 1 in 1947. The family income data are posttransfer-pretax. Inflation adjustment for each series is via the CPI-U-RS. Data sources: Bureau of Economic Analysis, “National Income and Product Accounts Tables,” table 1.1.5; Census Bureau, “Historical Income Data,” tables F-1 and F-5.

How much income has this cost middle-class households? To see, let’s make two adjustments. First, let’s switch from families to households. A family is defined by the Census Bureau as a household with two or more related persons. Defined this way, families don’t include adults who live alone or with others to whom they aren’t related. It’s odd to exclude this group, but that’s what the Census Bureau did until 1967. Only then did it begin tabulating data for all households. I used families in figures 4 in order to begin earlier, in the mid-1940s, but household is a better unit. Second, to make things perfectly comparable, let’s switch from GDP per person to GDP per household.6

As figure 5 shows, median household income was $50,000 in 1979 and $54,000 in 2014. Had it kept pace with GDP per household since 1979, it would instead have been $68,000.

Figure 5. Median household income
Posttransfer-pretax income, in 2015 dollars. Inflation adjustment is via the CPI-U-RS. “k” = thousand. Data sources: Bureau of Economic Analysis, “National Income and Product Accounts Tables,” table 1.1.5; Census Bureau, “Historical Income Data,” table H-5.

Why has this happened? Rising inequality. Since the 1970s, a larger and larger share of household income growth has gone to Americans at the very top of the ladder — roughly speaking, those in the top 1%. The income pie has gotten bigger, and everyone’s slice has increased in size, but the slice of the richest has expanded massively while that of the middle and below has gotten only a little bigger.7

This is a disappointing development. But does the trend in middle incomes paint an accurate picture of changes in living standards?8


In the view of some, the picture conveyed by figures 4 and 5 is too pessimistic. They argue that incomes or broader living standards actually have grown relatively rapidly, keeping pace with the economy.9 There are nine variants of this view. Let’s consider them one by one.

1. The income data are too thin. The data for family income in figures 4 and 5 don’t include certain types of government transfers or the value of health insurance contributions from employers or (in the case of Medicare and Medicaid) government. And they don’t subtract taxes. If these sources of income have risen rapidly for middle-class households, or if taxes have fallen sharply, the picture conveyed in figures 4 and 5 will understate the true rate of progress.

Happily, we have an alternative source of information: data compiled by the Congressional Budget Office (CBO). I didn’t use these data in figures 4 and 5 because they don’t begin until 1979. But if those charts are replicated using the CBO data, the trends since the late 1970s look similar.10

2. The income data miss upward movement over the life course. The family income data shown in figures 4 and 5 are from the Current Population Survey, which each year asks a representative sample of American adults what their income was in the previous year. But each year the sample consists of a new group; the survey doesn’t track the same people as they move through the life course.

If we interpret figures 4 and 5 as showing what happens to typical American families over the life course, we’ll conclude that they see very little increase in income as they age. But that’s incorrect. In any given year, some of those with below-median income are young. Their wages and income are low because they are in the early stages of the work career and/or because they’re single. Over time, many will experience a significant income rise, getting pay increases or partnering with someone who also has earnings, or both. Figures 4 and 5 miss this income growth over the life course.

Figure 6 illustrates the point. The lower line shows median income among families with a “head” aged 25 to 34. The top line shows median income among the same cohort of families twenty years later, when their heads are aged 45 to 54. Consider the year 1979, for instance. The lower line tells us that in 1979 the median income of families with a 25-to-34-year-old head was about $58,000 (in 2013 dollars). The data point for 1979 in the top line looks at the median income of that same group of families twenty years later, in 1999, when they are 45 to 54 years old. This is the peak earning stage for most people, and their median income is now about $91,000.

Figure 6. Median income within and across cohorts
For each year, the lower line is median income among families with a “head” age 25-34 and the top line is median income for the same cohort of families twenty years later. In the years for which the calculation is possible, 1947 to 1993, the average increase in income during this two-decade portion of the life course is $32,500. The data are in 2013 dollars; inflation adjustment is via the CPI-U-RS. “k” = thousand. Data source: Census Bureau, “Historical Income Data,” table F-11.

In each year, the gap between the two lines is roughly $33,000. This tells us that the incomes of middle-class Americans tend to increase substantially as they move from the early years of the work career to the peak years.

Should this reduce our concern about the over-time pattern shown in figures 4 and 5 above? No, it shouldn’t. Look again at figure 6. Between the mid-1940s and the mid-1970s, the median income of families in early adulthood (the lower line) rose steadily. In the mid-1940s median income for these young families was around $27,000; by the mid-1970s it had doubled. Americans during this period experienced income gains over the life course, but they also tended to have higher incomes than their predecessors, both in their early work years and in their peak years. That’s because the economy was growing at a healthy clip and the economic growth was trickling down to Americans in the middle.

After the mid-1970s, this steady gain disappeared. From the mid-1970s to 2013 the median income of families with a 25-to-34-year-old head was flat. They continued to achieve income gains during the life course. (Actually, we don’t yet know about those who started out after the mid-1990s, as they’re just now beginning to reach age 45 to 54. The question marks in the chart show what their incomes will be if the historical trajectory holds true.) But the improvement across cohorts that characterized the period from the mid-1940s through the 1970s — each cohort starting higher and ending higher than earlier ones — disappeared.

Income for many Americans rises during the life course, and this is indeed hidden by charts such as figures 4 and 5. But that shouldn’t lessen our concern about the decoupling of household income growth from economic growth that has occurred over the past generation. We want improvement not just within cohorts, but also across them.

3. Households have gotten smaller. The size of the typical American household has been shrinking since the mid-1960s, when the “baby boom” ended. Some will therefore say we don’t need income growth to be so rapid any more.

But this shrinkage in household size probably shouldn’t alter our interpretation of slow income growth. Incomes have become decoupled from economic growth because a large and rising share of economic growth has gone to households at the top of the ladder. Yet household size has decreased among the rich too; they don’t need the extra income more than those in the middle and below do.

4. More people are in college or retired. The income data in figures 4 and 5 are for families with a “head” aged 15 or older. The share of young Americans attending college has increased since the 1970s, and the share of Americans who are elderly and hence retired has risen. These two developments have reduced the share of families with an employed adult head. However, this doesn’t account for the slow growth of family income relative to the economy. The trend in income among families with a head aged 25 to 54, in the prime of the work career, is very similar to that for all families.11

5. There are more immigrants. Immigration into the United States began to increase in the late 1960s. The foreign-born share of the American population, including both legal and illegal immigrants, rose from 5% in 1970 to 13% in 2013.12 Quite a few have come with limited labor market skills and little or no English, so their incomes tend to be low. For many such immigrants, a low income in the United States is a substantial improvement over what their income would be in their home country. So if this accounts for the divorce between economic growth and median income growth over the past generation, perhaps we shouldn’t worry.

Immigration is indeed part of the story. But it is a relatively small part. The rise in median family income for non-Hispanic whites, which excludes most immigrants, has been only slightly greater than the rise in median income for all families shown in figures 4 and 5.13

6. Employer spending on healthcare and pensions for their employees has increased. This might leave less money available for wage and salary increases. However, the share of employee compensation that goes to nonwage benefits has been essentially flat since the late 1970s, so this has played at most a very small role in the failure of household incomes to keep up with per capita GDP during this period.14

7. Consumption has continued to rise rapidly. Some consider spending a better indicator than income of people’s standard of living. Even though the incomes of middle- and low-income Americans have grown slowly, they may have increased their consumption more rapidly by drawing on assets (equity in a home, savings) and/or debt.

But that isn’t the case. According to the best available data, from the Consumer Expenditures Survey (CES), median family expenditures rose at the same pace as median family income in the 1980s, 1990s, and 2000s.15

8. Wealth has increased sharply. Income and consumption growth for middle-income Americans may have lagged well behind growth of the economy, but was that offset by rapid growth of wealth (assets minus debts)?

Yes, it was, but only temporarily. We have data on wealth from the Survey of Consumer Finances (SCF), administered by the Federal Reserve every three years. Figure 7 shows the trend in median household wealth along with the trend in median household income. The wealth data are first available in 1983. What we see is a sharp upward spike in median wealth in the second half of the 1990s and the first half of the 2000s. The home is the chief asset of most middle-class Americans, and home values jumped during this period. But then the housing bubble burst and median wealth fell precipitously, erasing all of the gains.16 And for those who lost their home during the crash, things are worse than what’s conveyed by these data.

Figure 7. Median household income and median household wealth
2013 dollars; inflation adjustment is via the CPI-U. “k” = thousand. Median wealth: Household net worth, calculated as assets minus liabilities. Data source: Edward N. Wolff, “Household Wealth Trends in the United States, 1962-2013,” Working Paper 20733, National Bureau of Economic Research, 2014, table 1, using Survey of Consumer Finances data. Median income: Posttransfer-posttax household income, adjusted for household size and then rescaled to reflect a three-person household. Data source: Luxembourg Income Study.

In fact, even before the bubble burst, not everyone benefited. Of the one third of Americans who don’t own a home, many are on the lower half of the income ladder. For them, the rise in home values in the 1990s and 2000s did nothing to compensate for the slow growth of income since the 1970s.

9. There have been significant improvements in quality of life. The final variant of the notion that income data understate the degree of advance in living standards focuses on improvements in the quality of goods, services, and social norms. It suggests that adjusting the income data for inflation doesn’t do justice to the enhancements in quality of life that have occurred in the past generation.

Fewer jobs require hard physical labor, and workplace accidents and deaths have decreased. Life expectancy rose from 74 years in 1979 to 79 years in 2012. Cancer survival is up. Infant mortality is down. An array of new pharmaceuticals now help relieve various conditions and ailments. MRIs, CT scans, and other diagnostic tools have enhanced physicians’ ability to detect serious health problems. Organ transplants, hip and knee replacements, and lasik eye surgery are now commonplace. Violent crime has dropped to pre-1970s levels. Air and water quality are much improved.

We live in bigger houses; the median size of new homes rose from 1,600 square feet in 1979 to 2,600 in 2013. Cars are safer and get better gas mileage. Food and clothing are cheaper. We have access to an assortment of conveniences that didn’t exist or weren’t widely available a generation ago: personal computers, printers, scanners, microwave ovens, TV remote controls, digital video recorders, camcorders, digital cameras, five-blade razors, home pregnancy tests, home security systems, handheld calculators. Product variety has increased for almost all goods and services, from cars to restaurant food to toothpaste to television programs.

We have much greater access to information via the internet, Google, cable TV, travel guides, mapping apps and GPS, smartphones, and tablets. We have a host of new communication tools: cell phones, voicemail, email, Skype, Facebook, Twitter, Instagram. Personal entertainment sources and devices have proliferated: cable TV, high-definition televisions, home entertainment systems, the internet, MP3 players, CD players, DVD players, Netflix, satellite radio, video games.

Last, but not least, discrimination on the basis of sex, race, and more recently sexual orientation have diminished. For women, racial and ethnic minorities, and lesbian and gay Americans, this may be the most valuable improvement of all.

There is no disputing these gains in quality of life. But did they occur because income growth for middle- and low-income Americans lagged well behind growth of the economy? In other words, did we need to sacrifice income growth in order to get these improved products and services?

Some say yes, arguing that returns to success soared in fields such as high tech, finance, entertainment, and athletics, as well as for CEOs. These markets became “winner-take-all,” and the rewards reaped by the winners mushroomed. For those with a shot at being the best in their field, this increased the financial incentive to work harder or longer or to be more creative. This rise in financial incentives produced a corresponding rise in excellence — new products and services and enhanced quality.

Is this correct? To begin, consider the case of Apple and Steve Jobs. Apple’s Macintosh, iPod, iTunes, MacBook Air, iPhone, and iPad were so different from and superior to anything that preceded them that their addition to living standards isn’t likely to be adequately measured. Did slow middle-class income growth make this possible? Would Jobs and his teams of engineers, designers, and others at Apple have worked as hard as they did to create these new products and bring them to market in the absence of massive winner-take-all financial incentives?

It’s difficult to know. But Walter Isaacson’s comprehensive biography of Steve Jobs suggests that he was driven by a passion for the products, for winning the competitive battle, and for status among peers.17 Excellence and victory were their own reward, not a means to the end of financial riches. In this respect Jobs mirrors scores of inventors and entrepreneurs over the ages. So while the rise of winner-take-all compensation occurred simultaneously with surges in innovation and productivity in certain fields, it may not have caused those surges.

For a more systematic assessment, we can look at the preceding period — the 1940s, 1950s, 1960s, and early 1970s.18 In these years lower-half incomes grew at roughly the same pace as the economy and as incomes at the top. Did this squash the incentive for innovation and hard work and thereby come at the expense of broader quality-of-life improvements?

During this period the share of Americans working in physically taxing jobs fell steadily, as employment in agriculture and manufacturing was declining. Life expectancy rose from 65 in 1945 to 71 in 1973. Antibiotic use began in the 1940s, and open-heart bypass surgery was introduced in the 1960s.

In 1940, only 44% of Americans owned a home; by 1970 that jumped to 64%. Home features and amenities changed dramatically, as the following list makes clear. Running water: 70% in 1940, 98% in 1970. Indoor flush toilet: 60% in 1960, 95% in 1970. Electric lighting: 79% in 1940, 99% in 1970. Central heating: 40% in 1940, 78% in 1970. Air conditioning: very few (we don’t have precise data) in 1940, more than half in 1970. Refrigerator: 47% in 1940, 99% in 1970. Washing machine: less than half in 1940, 92% in 1970. Vacuum cleaner: 40% in 1940, 92% in 1970.

In 1970, 80% of American households had a car, compared to just 52% in 1940. The interstate highway system was built in the 1950s and 1960s. In 1970 there were 154 million air passengers, versus 4 million in 1940. Only 45% of homes had a telephone in 1945; by 1970 virtually all did. Long-distance phone calls were rare before the 1960s. In 1950, just 60% of employed Americans took a vacation; in 1970 that had risen to 80%. By 1970, 99% of Americans had a television, up from just 32% in 1940. In music, the “album” originated in the late 1940s, and rock-n-roll began in the 1950s. Other innovations that made life easier or more pleasurable include photocopiers, disposable diapers, and the bikini.

The Civil Rights Act of 1964 outlawed gender and race discrimination in public places, education, and employment. For women, life changed in myriad ways. Female labor force participation rose from 30% in 1940 to 49% in 1970. Norms inhibiting divorce relaxed in the 1960s. The pill was introduced in 1960. Abortion was legalized in the early 1970s. Access to college increased massively in the 1960s.

Comparing these changes in quality of life is difficult, but I see no reason to conclude that the pace of advance, or of innovation, has been more rapid in recent decades than before.19

Yes, there have been significant improvements in quality of life in the United States since the 1970s. But that shouldn’t lessen our disappointment in the fact that incomes have been growing far more slowly than the economy.


Rather than understating the true degree of progress for middle- and low-income Americans, the income trends shown in figures 4 and 5 above might overstate it.20

1. Income growth is due mainly to the addition of a second earner. The income of American households in the lower half has grown slowly since the 1970s. But it might not have increased at all if not for the fact that more households came to have two earners rather than one. From the 1940s through the mid-1970s, wages rose steadily. As a result, the median income of most families, whether they had one earner or two, increased at about the same pace as the economy.21 Since then, the earnings of Americans working full-time and year-round have barely budged.22

It’s important to emphasize that most of this shift from one earner to two has been voluntary. A growing number of women seek employment, as their educational attainment has increased, discrimination in the labor market has dissipated, and social norms have changed. The transition from the traditional male-breadwinner family to the dual-earner one isn’t simply a product of desperation to keep incomes growing.

Even so, the fact that income growth for lower-half households has required adding a second earner has two problematic implications. First, households with a single adult have seen no income rise at all.23 Second, as more two-adult households have both adults in employment, more are struggling to balance the demands of home and work. High-quality childcare and preschool are expensive, and elementary and secondary schools are in session only 180 of the 250 weekdays each year. The difficulty is accentuated by the growing prevalence of long work hours, odd hours, irregular hours, and long commutes. By the early 2000s, 25% of employed men and 10% of employed women were working 50 or more hours per week.24 And 35-40% of Americans were working outside regular hours (9 am to 5 pm) and/or days (Monday to Friday).25 Average commute time rose from 40 minutes in 1980 to 50 minutes in the late 2000s.26

2. The cost of key middle-class expenses has risen much faster than inflation. The income numbers in figures 4 and 5 are adjusted for inflation. But the adjustment is based on the price of a bundle of goods and services considered typical for American households. Changes in the cost of certain goods and services that middle-class Americans consider essential may not be adequately captured in this bundle. In particular, because middle-class families typically want to own a home and to send their kids to college, they suffered more than other Americans from the sharp rise in housing prices and college tuition costs in the 1990s and 2000s. Moreover, as middle-class families have shifted from having one earner to two, their spending needs may have changed in ways that adjusting for inflation doesn’t capture. For example, they now need to pay for childcare and require two cars rather than one.27

Consider a four-person family with two adults and two preschool-age children. In the early 1970s, this family would probably have had one of the adults employed and the other staying at home. By the mid-2000s, it’s likely that both were employed. Here is how their costs for these big-ticket expenses might have differed.28 Childcare: $0 in the early 1970s, $12,500 in the mid-2000s. Car(s): $5,800 for one car in the early 1970s, $8,800 for two cars in the mid-2000s. Home mortgage: $6,000 in the early 1970s, $10,200 in the mid-2000s. When the children reach school age, the strain eases. But when they head off to college it reappears; the average cost of tuition, fees, and room/board at public four-year colleges rose from $6,500 in the early 1970s to $12,000 in the mid-2000s.29


Figure 8 shows change in median household income by change in GDP per capita in the US and 13 other countries. Median income increased less in the United States than in most of the other nations. That’s not because our economy grew less rapidly; in fact, our increase in per capita GDP was comparatively large. The problem is that less of America’s economic growth reached middle-class households. (Figure A2 in the appendix has graphs that show the pattern for each country.)

Figure 8. Median household income growth by economic growth
Change is per year, 1979-2013. Because the actual years vary somewhat depending on the country, change is calculated by regressing household income or GDP per capita on year. Household incomes are posttransfer-posttax, adjusted for household size (the amounts shown are for a household with three persons). Household incomes and GDP per capita are adjusted for inflation and converted to US dollars using purchasing power parities. Ireland and Norway are omitted; both would be far off the plot in the upper-right corner. Data sources: OECD; Luxembourg Income Study.

The reason, again, is income inequality. As figure 9 shows, in countries where the top 1% get a larger share of the income, median income (adjusted for economic growth) has tended to grow more slowly. Though it’s an especially striking case, the United States is part of a widespread pattern.

Figure 9. Median household income growth by top 1%’s income share
1979-2013. The line is a linear regression line. The vertical axis measure is the residuals from a regression of change in median household income on change in GDP per capita. The incomes are adjusted for household size and then rescaled to reflect a three-person household, adjusted for inflation, and converted to US dollars using purchasing power parities. Top 1% income share: Income share of the top 1% of tax units. Pretax income, excluding capital gains. Source: Lane Kenworthy and Timothy Smeeding, “Does Top-End Income Inequality Reduce Middle-Class Income Growth?,” 2015, using data from the World Top Incomes Database, the Luxembourg Income Study, and the OECD.


Is the rise in top-end income inequality in the US since the late 1970s a function of employment shifts? No. Households in the top 1% often have two employed adults, but so do many middle-income households.30

Is it because the share of value-added in the economy going to labor fell dramatically? The labor share did fall, but the decline was fairly minor, and smaller than in many other rich nations.31

The main obstacle to income growth for middle-class households has been rising inequality of wages. As figure 10 shows, wages for those in the top 1% have increased sharply since the late 1970s, while in the middle they have grown very slowly.32 And of the rich nations for which we have data on wage trends, the United States has had the slowest growth at the median.33

Figure 10. GDP per capita and wages
Each series is displayed as an index set to equal 1 in 1947. Inflation adjustment for each series is via the CPI-U-RS. Data sources: Bureau of Economic Analysis, “National Income and Product Accounts Tables,” table 1.1.5; Council of Economic Advisers, Economic Report of the President; Lawrence Mishel et al, The State of Working America, 12th edition, wages dataset.

What caused this divergence in wage trends? At the moment, social scientists have many hypotheses and few solid answers. The suspects include:

  • More growth of demand than supply for highly-skilled labor and more growth of supply than demand for less-skilled workers
  • Increase in product market competition
  • Technological advance
  • Globalization
  • Shift from manufacturing to services
  • Union decline
  • Stagnant statutory minimum wage
  • Emergence of winner-take-all markets
  • More aggressive shareholders
  • Changes in CEO compensation practices: greater use of stock options, collusion among directors
  • Rising stock values
  • Deregulation of finance


Since the 1970s, incomes have risen slowly for the broad middle of American households, despite sustained growth in the economy. With the top 1% getting a larger and larger portion, household income growth for the middle has become decoupled from economic growth. America’s middle class is still fairly well-off by comparative and historical standards, but it could be doing a good bit better.


The appendix has additional data.

  1. John Rawls, A Theory of Justice, Harvard University Press, 1971; Lane Kenworthy, Progress for the Poor, Oxford University Press, 2011. 
  2. I exclude tiny Luxembourg. 
  3. Data on household income at the median are available from not only the Luxembourg Income Study but also the OECD, so more countries and years are included. 
  4. See also David Leonhardt and Kevin Quealy, “The American Middle Class Is No Longer the World’s Richest,” New York Times, 2014. 
  5. I use the 20th and 80th percentiles here rather than the 25th and 75th because of data availability. 
  6. I use households rather than persons as the denominator for GDP because the number of households has increased faster than the number of persons since the late 1970s. Note also that I use the same price deflator for both GDP and median income. See Brian Nolan, Max Roser, and Stefan Thewissen, “GDP Per Capita Versus Median Household Income: What Gives Rise to Divergence Over Time?,” Working Paper 672, Luxembourg Income Study, 2016. 
  7. Lane Kenworthy, “Income Inequality,” The Good Society. 
  8. The following two sections draw from Lane Kenworthy, Social Democratic America, Oxford University Press, 2014, ch. 2. 
  9. See, for instance, Robert J. Samuelson, The Good Life and Its Discontents, Times Books, 1995; W. Michael Cox and Richard Alm, Myths of Rich and Poor, Basic Books, 1999; Gregg Easterbrook, The Progress Paradox, Random House, 2003; Stephen J. Rose, Rebound, St. Martin’s Press, 2010; Richard V. Burkhauser, Jeff Larrimore, and Kosali I. Simon, “A ‘Second Opinion’ on the Economic Health of the American Middle Class,” Working Paper 17164, National Bureau of Economic Research, 2011; Bruce D. Meyer and James X. Sullivan, “The Material Well-Being of the Poor and the Middle Class Since 1980,” Working Paper 2011-04, American Enterprise Institute, 2011. 
  10. Lane Kenworthy, “Is Decoupling Real?,” Consider the Evidence, March 12, 2012. 
  11. Actually, it’s worse than the trend for all families. That’s because median income among the retired has been growing at a slightly faster pace than median income among prime-working-age families, due to rising Social Security benefit levels. The data are at Census Bureau, “Historical Income Data,” table F-11. 
  12. Census Bureau, “Foreign Born.” 
  13. The data are at Census Bureau, “Historical Income Data,” table F-5. 
  14. Jess Bailey, Joe Coward, and Matthew Whittaker, “Painful Separation: An International Study of the Weakening Relationship between Economic Growth and the Pay of Ordinary Workers,” Commission on Living Standards, Resolution Foundation, 2011, figure A3, using data from the OECD; Lawrence Mishel, Josh Bivens, Elise Gould, and Heidi Shierholz, The State of Working America, 12th edition, Economic Policy Institute, 2012, pp. 180-183, using data from the Bureau of Labor Statistics Employer Costs for Employee Compensation (ECEC) survey. 
  15. David S. Johnson, “Using Expenditures to Measure the Standard of Living in the United States: Does It Make a Difference?,” in What Has Happened to the Quality of Life in the Advanced Industrialized Nations?, edited by Edward N. Wolff, Edward Elgar, 2004; Meyer and Sullivan, “The Material Well-Being of the Poor and the Middle Class Since 1980.” See also Orazio Attanasio, Erik Hurst, and Luigi Pistaferri, “The Evolution of Income, Consumption, and Leisure Inequality in the US, 1980-2010,” Working Paper 17982, National Bureau of Economic Research, 2012. 
  16. See Edward N. Wolff, “The Asset Price Meltdown and the Wealth of the Middle Class,” Russell Sage Foundation, 2012. 
  17. Walter Isaacson, Steve Jobs, Simon and Schuster, 2011. 
  18. The data in the following paragraphs come from a variety of sources, including Stanley Lebergott, The American Economy, Princeton University Press, 1976; Cox and Alm, Myths of Rich and Poor; Easterbrook, The Progress Paradox; Claude S. Fischer, Made in America, University of Chicago Press, 2010. 
  19. For an argument that the pace of innovation has been less rapid since the mid-1970s, see Tyler Cowen, The Great Stagnation, Penguin, 2011. 
  20. See in particular Elizabeth Warren and Amelia Warren Tyagi, The Two-Income Trap, Basic Books, 2003; Monica Lesmerises, “The Middle Class at Risk,” Century Foundation, 2007; Rebecca M. Blank, “Middle Class in America,” US Department of Commerce, 2010. 
  21. Census Bureau, “Historical Income Data,” table F-12. 
  22. Census Bureau, “Historical Income Data,” table P-38. 
  23. Census Bureau, “Historical Income Data,” table F-12. 
  24. Jerry A. Jacobs and Kathleen Gerson, The Time Divide, Harvard University, 2004. 
  25. Harriet Presser, Working in a 24/7 Economy, Russell Sage Foundation, 2003. 
  26. Brian McKenzie and Melanie Rapino, “Commuting in the United States: 2009,” US Census Bureau, 2011. 
  27. Some suggest that the cost of goods and services consumed by low-income households has risen less rapidly than that of the bundle used to adjust for inflation. See Christian Broda and John Romalis, “Inequality and Prices: Does China Benefit the Poor in America?,” unpublished paper, 2008. Others argue that this is wrong. See Meyer and Sullivan, “The Material Well-Being of the Poor and the Middle Class Since 1980.” 
  28. Lesmerises, “The Middle Class at Risk,” figure 12, using data from Warren and Tyagi, The Two-Income Trap. See also J. Bradford DeLong, “The Changing Structure of Prices since 1960,” Grasping Reality, December 8, 2012. 
  29. College Board, “Trends in College Pricing 2006,” table 4a. 
  30. Since 1979, the share of working-age American households with two (or more) employed persons always been in the neighborhood of 55-65%. See Census Bureau, “Historical Income Data,” tables H-10 and H-12. 
  31. Bailey et al, “Painful Separation,” figure A2, using OECD data. 
  32. Lawrence Mishel and colleagues (The State of Working America, 12th edition, p. 311) explain the calculation of the wage data for the top 1% and the bottom 90% as follows: “Data are taken from Kopczuk, Saez, and Song (“Earnings Inequality and Mobility in the United States: Evidence from Social Security Data Since 1937,” Quarterly Journal of Economics, 2010, table A-3). Data for 2006 through 2010 are extrapolated from 2004 data using changes in wage shares computed from Social Security Administration (SSA) wage statistics (…). 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.” 
  33. Bailey et al, “Painful Separation,” figure A1, using OECD data.