T H E M E R I T O C R A C Y T R A P How America’s Foundational Myth Feeds Inequality, Dismantles the Middle Class, and Devours the Elite DANIEL MARKOVITS Copyright © 2019 by Daniel Markovits Figures and Tables f i g u r e 1 Average Hours Worked per Week by Income Rank (Ten-Year Moving Averages) FIGU R E 1 traces the association between income and industry over the past three-quarters of a century. Workers in the bottom 60 percent of the income distribution work nearly ten fewer hours per week today than they did in 1940, a decline of about 20 percent. Workers in the next 30 percent of the distribution (who lie between the 60th and the 90th percentiles) have worked effectively constant hours over this period. Work hours in the top tenth of the distribution, by contrast, have increased, with growing increases as incomes rise into the narrow elite. The top 1 percent in particular increased its work hours by nearly seven per week, which is more than any lower-income cohort. Uniquely, one-percenters also continued to increase their work hours even in the 2000s. The cumulative effects of this trend are enormous. At midcentury, one-percenters worked between three and four hours per week less than workers in the bottom 60 percent. To- day, they work roughly twelve hours per week more. The two components of this realignment (with the 1 percent working twelve hours more per week rather than three to four hours less per week) cumulate to roughly sixteen work hours, or two regulation workdays, per week. Finally, these numbers—because they report on only full-time, non-self-employed, prime-aged men— almost certainly understate the actual trends. Most important, they do not take into account trends in unemployment and especially labor force participation, which again shift work effort away from the middle class and toward the elite. 1950 1960 1970 1980 1990 2000 2010 99–100 Income percentile 90–99 60–90 0–60 35 37 39 41 43 45 47 49 51 53 55 1940 average weekly hours worked f i g u r e 2 Income Poverty, Consumption Poverty, and the Income Share of the Top 1 Percent (Five-Year Moving Averages) FIGU R E 2 shows trends in income poverty and consumption poverty (on the left axis) and the income share of the top 1 percent (on the right axis) from 1960, at the center of the Great Com- pres sion, through the new millennium. The two solid gray lines that concern poverty both slope down. Although precise trends depend on how one counts, poverty has fallen to between about one-half and one-sixth of its 1960 levels. Income poverty rates have decreased from about 22.5 per- cent to about 12 percent. Consumption poverty rates have decreased from about 31 percent to less than 5 percent. By contrast, the dashed black line that concerns wealth slopes steeply up: the best-off 1 percent have roughly doubled their share of economic advantage since 1960—reflecti ng an abso- lute increase in the top 1 percent’s income share from about 10 percent to about 20 percent. 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015 poverty rate Income poverty Consumption poverty Top 1% income share 0.0% 5.0% 10.0% 15.0% 20.0% income share f i g u r e 3 Ratios of Representative High, Middle, and Low Incomes over Time (Five-Year Moving Averages) 0 5 10 15 20 25 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 2014 Ratio of average income of 1% to median income Ratio of median income to average income of bottom 20% FIGU R E 3 displays trends in the ratios of post-tax incomes at key points in the overall income distribution. The dark dashed line that slopes upward reports that the ratio of the average income of the top 1 percent to the income of the middle class (defined as the 50th percentile) has risen. The rich, that is, are getting richer relative to the middle class—they are leaving the middle class behind—and the average one-percenter today captures more than twenty times the median in- come, or nearly twice the multiple of his counterpart in the 1960s and 1970s. The light solid line reports the ratio of the median income to the average income of the poorest 20 percent. The line’s slight downward slope overall reveals that the median earner captures a little less income relative to the poor today than at midcentury—that the poor and the middle class are converging. f i g u r e 4 U.S. Top-End, Bottom-End, and Full Gini Coefficients over Time (Five-Year Moving Averages) 0.20 0.25 0.30 0.35 0.40 0.45 0.50 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 2012 2016 gini coefficient Full Gini Top 5% Bottom 70% FIGU R E 4 shows the Gini coefficients of the United States, calculated in three ways. The upward- sloping dark gray line displays the Gini for the entire U.S. economy. Its steep rise reflects the commonplace sense that inequality has shown a stark increase, from levels that resembled Nor- way in 1964 to levels that resemble India today. The light gray line is less familiar. It displays the Gini index for the bottom 70 percent of the U.S. income distribution, constructed not by redistributing any income but simply by discarding all income from the top 30 percent of house- holds. The figure reveals that this bottom-end Gini has fallen (by about 10 percent) since mid- century, so that there has been a modest decrease in inequality across the bottom seven-tenths of the U.S. income distribution. Finally, the dashed black line with the steepest upward slope rep- resents the Gini for the top 5 percent of the income distribution, now constructed by discarding all the income from the bottom 95 percent. This line shows that inequality within the rich has skyrocketed. Moreover, the difference between inequality within the large bottom and within the narrow top—the gap between black dashed and light gray lines—remained roughly steady between 1964 and 1984 but increased sharply beginning in 1984. Economic inequality’s center of gravity is moving up the income distribution. Indeed, the black dashed and the dark gray lines have recently crossed: inequality within the rich now exceeds inequality in the overall economy, a result that would have been unimaginable at midcentury, when the central economic divide separated the poor from the middle class. f i g u r e 5 Ratios of Education Expenditures by Income and Education (Five-Year Moving Averages) 0 1 2 3 4 5 6 7 8 9 1984 1988 1992 1996 2000 2004 2008 2012 ratio of education expenditure by income status Ratio of education expenditure of top-income quintile to middle-income quintile Ratio of education expenditure of middle-income quintile to second-lowest-income quintile Ratio of education expenditure of postgraduate degree holders to those with only a high school degree Ratio of education expenditure of high school graduates to those with no high school degree 0 2 4 6 8 10 12 14 ratio of education expenditure by education status FIGU R E 5 displays trends in the ratios of consumption expenditures specifically on education between rich and middle-class households on the one hand and between middle-class and poor households on the other. The figure reveals a massive increase in the investments that rich house- holds make in children’s education, relative to the investments made by the middle class. At the same time, investments made by middle-income households have not increased relative to invest- ments made by poor households. A second, much briefer series—which reports ratios of educa- tion expenditures between super-educated and ordinarily educated households and between ordinarily educated and uneducated households—confirms the lesson of the first. The education series also selects out a narrower elite than the income series and, strikingly, reveals still more disproportionate educational investment, compared to the middle class. Note the close correspondence between the expenditure ratios and the income ratios re- ported in Figure 3. In each case, a relatively stable midcentury order, in which the principal in- equalities concerned differences between the middle class and the poor, gives way (beginning sometime in the 1980s) to a new order, in which the top separates itself from the middle, even as the middle and the bottom slowly converge. f i g u r e 6 90/50 and 50/10 Income Achievement Gaps for Reading and Math average difference in standardized test scores Trends in 90/50 and 50/10 Income Gaps in Reading, 1943–2001 Cohorts 1.00 Fitted trend 50/10 gap Fitted trend 90/50 gap 0.75 0.50 0.25 0.00 1940 1950 1960 1970 1980 1990 2000 cohort birth year Trends in 90/50 and 50/10 Income Gaps in Math, 1943–2001 Cohorts average difference in standardized test scores 1.00 0.75 0.50 0.25 0.00 1940 1950 1960 1970 cohort birth year 1980 1990 2000 Fitted trend 50/10 gap Fitted trend 90/50 gap FIGU R E 6 (Top and Bottom), constructed by the sociologist Sean Reardon, reports the school achievement gaps for reading (top) and math (bottom) between the 90th and 50th percentiles of the income distribution on the one hand, and between the 50th and 10th percentiles on the other. This exercise reveals that the 90/50 gaps have been rising since midcentury and have been rising increasingly steeply since the early 1970s. The 50/10 gaps, by contrast, have been rising much more slowly and (for reading) have even begun to decline. The combined effect of the two trends entails that while at midcentury the 50/10 gaps roughly doubled the 90/50 gaps in reading and were about a third greater in math, by the mid-1990s the 90/50 gaps had caught up. More- over, the 90/50 gaps have continued growing since, even as the 50/10 gaps have leveled off and even begun to fall. Today the school achievement gaps between rich and middle-class children are between a quarter and a third greater than the gaps between the middle class and the poor. Note again the close correspondence between the achievement gaps and the income ratios reported in Figure 3. In each case, a relatively stable midcentury order, in which the principal inequalities concern differences between the middle class and the poor, gives way (beginning sometime in the 1980s) to a new order, in which the top separates itself from the middle, even as the middle and the bottom slowly converge. f i g u r e 7 GDP Share, Employment Share, and Relative Income and Education for Finance, 1947–2005 (Five-Year Moving Averages) –0.05 0.05 0.15 0.25 0.35 0.45 0.55 0.0% 1.0% 2.0% 3.0% 4.0% 5.0% 6.0% 7.0% 8.0% 1947 1950 1953 1956 1959 1962 1965 1968 1971 1974 1977 1980 1983 1986 1989 1992 1995 1998 2001 2004 2007 ratio (income) rescaled difference (education) GDP share (left axis) Employment share (left axis) Relative education (right axis) Relative income (right axis) FIGU R E 7 shows two pairs of trends—output and employment on the left axis and relative in- come and education on the right—in the finance sector over the past seventy years. From the end of the Second World War through the end of the 1970s, finance was a mid-skilled industry that grew by hiring more workers. Finance’s shares of GDP and total employment grew together during this period. Moreover, and in line with their average productivity, finance-sector workers in this period were not appreciably better educated or paid than their private-sector counterparts. Then, from the 1980s onward, finance’s share of GDP accelerated its growth, even as finance’s employment share flattened and indeed began gently to decline. Furthermore, finance workers’ rising relative productivity (as relatively fewer workers accounted for relatively more GDP) was unsurprisingly accompanied by their increasing relative education and relative income. Note, al- though the series is not included in the figure, that finance’s share of total compensation paid rose steadily alongside its GDP share throughout both periods. Finance workers, in other words, did not take a bigger cut of their product. Instead, they became increasingly highly paid because they divided a stable slice of a growing pie among relatively fewer increasingly elite workers. f i g u r e 8 Percent Changes in Employment Shares for Routine and Fluid Skills 8% –5% 1% 11% –7% 1% 15% –14% 9% –20% –15% –10% –5% 0% 5% 10% 15% 20% 1982–1992 1992–2002 2002–2017 Routine Nonroutine cognitive Nonroutine manual FIGU R E 8, constructed by the economists Nir Jaimovich and Henry Siu, shows that each of the past three decades has seen a flight from mid-skilled routine-intensive work combined with a modest increase in manual, largely low-skilled fluid work and a massive increase in cognitive, high-skilled fluid work. Altogether, almost a quarter of the economy’s mid-skilled jobs have dis- appeared since 1980, and the share of jobs allocated specifically and exclusively to high-skilled workers has increased by nearly a third. f i g u r e 9 Earnings Segmentation by Education Level (Smoothed) 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% percent who earn at least lifetime earnings (smoothed) % No high school earning at least % High school earning at least % Bachelor’s earning at least % Professional earning at least Bachelor’s median ($2.27M) Professional median ($3.65M) No high school median ($0.97M) High school only median ($1.3M) $5,000,000 FIGU R E 9 shows the segmentation of income according to education. The completeness of the segmentation amazes. Only 7.3 percent of workers without a high school degree and only 14.3 percent of workers with a high school degree only earn as much as the median college graduate. Just 1.3 percent of high school dropouts, just 2.4 percent of high-school-only workers, and just 17.2 percent of workers with a BA only earn as much as the median professional school graduate. These numbers reveal that uneducated and educated workers live in almost entirely separate worlds, which effectively never overlap. The least educated face a constant, demoralizing struggle to find work at all, while (contrary to popular stories of college graduates living in their parents’ basements) the most educated enjoy full employment. And when they do find jobs, only about one worker in fifty from the bottom half of the educational distribution earns as much as the median worker from the top twentieth. f i g u r e 1 0 Incomes of the Bottom 90 Percent and Per Capita Consumption and Debt over Time (Ten-Year Moving Averages) $0 $5,000 $10,000 $15,000 $20,000 $25,000 $30,000 $35,000 $40,000 $45,000 $50,000 1947 1954 1961 1968 1975 1982 1989 1996 2003 2010 Per capita household debt Per capita consumption Bottom 90% average income FIGU R E 10 shows per capita consumption, household debt, and mean income for households in the bottom 90 percent of the income distribution, from 1947 through 2010. Consumption in- creased remarkably steadily over the seven decades depicted. The trends for income and borrow- ing, by contrast, each display a marked kink, and the kinks form a sort of mirror image. Mean income among the bottom 90 percent rose steadily (more or less in tandem with consumption) between 1947 and roughly 1975, at which point income stopped rising almost completely, even as consumption continued its smooth growth. Mean debt, by contrast, rose more slowly than rising incomes between 1947 and roughly 1975 and then, just a few years after incomes stopped growing, began a steep rise (more or less in tandem with still-rising consumption). The pattern is unmistakable: rising middle-class standards of living were once financed by growing middle- class incomes; then, beginning about 1975 and running through the present day, income stag- nated and borrowing rose sharply. In the face of rising market inequality, the United States financed the lifestyle of its middle class not through redistribution, but rather through debt. Borrowing propped up consumption as income fell short. Credit issued to middle-class house- holds is increasingly a thrift good, provided in the shadow of economic inequality and on the same basic model as payday lending. f i g u r e 1 1 The Returns to Skill and Unequal Investments in Education Australia Austria Canada Czech Republic Denmark Estonia Finland Germany Ireland Italy Japan South Korea New Zealand Norway Poland Slovakia Spain Sweden United States England 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 10 15 20 25 30 literacy gap given parents’ education tertiary wage premium FIGU R E 11 reports the relationship between the returns to skill and the inequality of educational investments across the OECD. The vertical axis displays the college wage premium, a straight- forward measure of the economic returns to worker skill as measured by the ratio of the median hourly wages of workers with and without college degrees. The horizontal axis displays the effect of parents’ status on children’s skill (measured by an international test of facility in “under- standing, using, reflecting on and engaging with written texts, in order to achieve one’s goals, develop one’s knowledge and potential, and participate in society”). Because parents’ education correlates highly with investment in children’s education, this is an excellent proxy for the degree of training concentration that a society produces. The striking correlation between the tertiary wage premium and the effect of parents’ education on children’s skill reveals that skill fetishism and training concentration vary across countries, not separately but together. f i g u r e 1 2 Children’s Changing Odds of Earning More Than Their Parents 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% –10% –20% –30% –40% –50% –60% –70% –80% –90% –100% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 1940 cohort 1980 cohort Decline in absolute mobility, 1940–1980 parent income percentile (conditional on positive income) percent of children earning more than their parents decline in absolute mobility, 1940–1980 FIGU R E 12 displays the percentage of children who earn more than their parents at midcentury and today, according to parents’ income ranks. The dashed light gray line tracks this measure of social mobility at midcentury, for children born in 1940. Strikingly, virtually all of these children—right across the income distribution—came to earn more than their parents, with the only exception being children of the very top earners, for whom this test inevitably established a high bar. The dashed dark gray line tracks the same measure for children born in 1980. It is lower everywhere, simply on account of slower economic growth in recent decades. But it also has a very different shape. For children born in 1980, the odds of earning more than their parents fall swiftly as parental income ranks rise to escape poverty and then more or less plateau, until they fall swiftly again for children of the very richest parents (again on account of the high earnings bar that these parents set). Finally, the solid black line casts meritocratic inequality’s effect on absolute mobility into sharp relief. The line reports the decline in absolute mobility across the two cohorts, again for every parental income rank. The decline is by far biggest for children whose parents’ incomes fell between roughly the 20th and roughly the 95th percentiles of the income distribution—that is, for the (very) broad middle class whom wage stagnation has hit hardest.