On May 8^th, Brookings officially launched a new initiative on the Future of the Middle Class. Through this initiative, we will publish research, analysis, and insights that are motivated by a desire to improve the quality of life for those in America’s middle class and to improve upward mobility into its ranks. We have already wrestled with how we define this group, considered its changing racial composition, and called upon experts to outline major policies geared toward improving its fate. But why all of this attention? Here are seven of the reasons we are worried about the American middle class.

1. Middle-class incomes are stagnant

Despite gains in national income over the past half-century, American households in the middle of the distribution have experienced very little income growth in recent decades:

The stagnant growth in median incomes is in stark contrast to the income trajectories of those at the very top of the income distribution. According to the Congressional Budget Office, the middle three household income quintiles experienced income growth rates of 28 percent from 1979 to 2014 in real terms. The top 20 percent – a group Richard Reeves refers to as “Dream Hoarders” – saw their incomes grow by 95 percent over the same period. Taxes and transfers reduce the level of inequality, but have done little to affect the trend. Further, the redistributive role of such policies is weaker in the U.S. than in other advanced economies, who have experienced smaller increases in income inequality than the U.S.

Trends in wealth inequality are even more dramatic than for income. One widely cited estimate from Emmanuel Saez and Gabrial Zucman suggests that the top 0.1 percent wealth share increased from 7 percent in 1978 to 22 percent in 2012.

Not only has income growth slowed in recent decades, measures of overall welfare that include the value of rising life expectancy and the impact of growing consumption inequality have slowed as well – at least since 2007 – as shown by our colleagues Ben Bernanke and Peter Olson in the first chapter of Brookings Big Ideas for America, edited by Michael O’Hanlon.

2. employment and wages are declining

The retreat from work among less-educated prime-age adults, especially men, is well-documented. One reason for these declines in employment and labor force participation is that work is less rewarding. Wages for those at the bottom and middle of the skill and wage distribution have declined or stagnated. The “college wage premium” has flattened in recent years, but workers with a college degree can still expect to earn over 80 percent more than those with just a high school diploma.

These trends are not only affecting those at the bottom – they are affecting workers in the middle as well. Our colleague Alan Berube shows that while workers of all skill levels were hard-hit by the Great Recession, those in the middle of the skill distribution (those with some college but no four-year degree) saw the steepest earnings declines.

Such declines predate the Great Recession. While wages have been starting to creep back up, lifetime earnings have been falling for decades for male workers. The lifetime incomes of men were up to one fifth lower for the cohort entering the labor market in 1983 compared to those who began work in 1967. That’s equivalent to about $136,000 over an assumed 30-year working life:

The story is the reverse for women, who are out-earning previous female cohorts. But they have yet to catch up with men. For both men and women, inequality in lifetime earnings is increasing.

3. Middle class children’s prospects are declining

Stagnant incomes and falling wages have meant that fewer Americans are growing up to be better off than their parents. Upward absolute intergenerational mobility was once the almost-universal experience among America’s youth. No longer. Among those born in 1940, about 90 percent of children grew up to experience higher incomes than their parents, according to researchers at the Equality of Opportunity Project. This proportion was only 50 percent among those born in the 1980s:

Whatever the rate of upward mobility, family background continues to play a strong role in determining who is upwardly mobile and who is not. Children born into poor families have a hard time making it into the middle class as adults, more so in the U.S. than in some other advanced countries such as Canada. Further, there appears to be a “glass floor” at the very top, protecting children from falling down the income ladder. These findings belie the very American idea that mobility is based only on hard work and merit. As class gaps in family formation, parenting styles, test scores, neighborhood quality, and college attendance and graduation continue to grow, the upward mobility of today’s children is increasingly at risk.

4. DESTINIES are diverging, especially by race

While the mobility and income trends reported above are concerning enough, they mask an even more troubling story for Black and Hispanic households:

The good news is that “the percentage of black households with incomes of at least $75,000 more than doubled from 1975 to 2016, adjusting for inflation” as William Julius Wilson reports. Across white, black, and Hispanic households, the share with incomes between $15,000 and $75,000 decreased over this period, and this was largely driven by increases in the share making more than $75,000. The share of households making less than $15,000 shrunk for each group. Still, gains were larger for white and Hispanic households:

As a result of these gains, the middle class is becoming more “race-plural,” in the words of Richard Reeves and Camille Busette. But will it remain racially diverse? Recent work by Raj Chetty and his colleagues shows that while Hispanics have relatively high rates of upward mobility, black Americans are less likely to move up the income ladder and more likely to fall down it. This trend is largely explained by the remarkably high rates of downward mobility among black men. Among those born into the middle-income quintile, over half of black boys fall into the bottom two quintiles as adults, compared to 40 percent of Hispanics and just over one-third of white men:

Downward mobility is much less common for black girls, suggesting a complicated interaction between being both black and male.

These mobility gaps by race can be attributed to various sources, including poverty, family structure, social capital, segregation and racial discrimination. As the recent Chetty et al. paper documents, neighborhoods with lower poverty rates, lower levels of racial bias, and higher rates of fatherhood presence among blacks tend to have smaller black-white mobility gaps. But these neighborhoods are few in number. (Racial bias was measured by scores on implicit bias tests and indices of explicit racial bias expressed in Google searches.)

5. place matters more than ever

Place matters. Where you end up depends a great deal not just on your family background and your race but also on where you grew up. The chance that a child born in the bottom quintile will make it to the top quintile as an adult ranges from around 4 percent in Charlotte to 13 percent in San Jose.

Alongside the diverging destinies of individuals is a great divergence in the prosperity of whole communities and regions of the country. Employment and economic growth is far from consistent from one metropolitan or rural area to the next. Because various parts of the country are home to different industries and occupations, trade and technology have had differential impacts by region. Employment rates have, for example, fallen most dramatically in the nation’s rural areas, though there are still more non-working men in cities than in the heartland.

Our Brookings colleagues Mark Muro and Jacob Whiton find that the largest metropolitan areas have accounted for the vast majority of the nation’s population, employment, and output growth since 2010. For the first time, the rural population is actually shrinking. This, alongside the declining role of manufacturing and mining, means that employment growth, an important contributor to overall economic growth, is waning in many parts of the country:

As Muro and Whiton put it, “The nation’s bigger communities…are now growing notably faster and accounting for more and more of the nation’s growth than before, even as small metros wane and most of the rural hinterland slides into deep decline.” As such, the destinies and future of middle-class households varies sharply by geography, and national income trends are not reflective of the experience in many communities. Some “booming” cities have experienced gains in median income while others, particularly older manufacturing communities in the Midwest and Southeast, have experienced sharp declines.

6. a sense of well-being has eroded, at least for some

Most Americans are unlikely to pore over the latest intergenerational mobility statistics. But many are acutely aware that the American Dream is a more distant prospect for today’s children. The share of Americans who are confident life for today’s children will be better than it has been for today’s adults was only 21 percent in 2014, down from one-third in 2007, according to a survey conducted by NBC News and the Wall Street Journal. Three-quarters are not confident today’s children will grow up to be better-off than their parents:

Similarly, according to Pew, more Americans think that today’s children will be financially worse off than their parents than those who believe they will be better off.

This lack of optimism for the future is echoed in other signs of social distress. As our colleague Carol Graham documents in her book Happiness for All?, hope, aspirations, and optimism vary considerably across the population. In fact, blacks and Hispanics are generally more optimistic than whites, conditional on income. This points to the importance of relative status. Black and Hispanic Americans might be materially disadvantaged compared to whites, but they have experienced at least some upward movement over the past several decades, while many poor and middle-class whites might feel that their relative status in society has actually declined over the same time period.

The evidence that social wellbeing is declining among many Americans, particularly whites, extends beyond the survey data. As Anne Case and Angus Deaton document, midlife mortality rates among non-Hispanic whites without college degrees is actually rising, while other educational and racial groups have experienced continued improvements in life expectancy and reduced midlife mortality. The “deaths of despair” that are the major contributors to this uptick in mortality among non-college whites include drug overdoses, suicides, and alcohol-related liver mortality. Though it is difficult to tease out cause and effect, the authors hypothesize that declining labor market opportunities and life prospects could be underlying this rise in mortality. So the declining economic prospects discussed above don’t just matter for intergenerational income mobility, they may also affect the emotional and social wellbeing of the middle class, and often in unpredictable ways.

7. middle-class families are more fragile and more dependent on two incomes

Though modest, the improvement in middle-class family incomes over the past several decades is entirely thanks to women’s added work hours and earnings. According to Heather Boushey and Kavya Vaghul at the Washington Center for Equitable Growth, were it not for women’s economic contributions, middle-income families would have experienced stark declines in income over the 1979 to 2013 period. The Council of Economic Advisers reported that the economy is 13.5 percent larger ($2 trillion) as a result of women’s increased participation and working hours since 1970. As they note, “Essentially all of the income gains that middle-class American families have experienced since 1970 are due to the rise in women’s earnings.”

However, women’s participation in the labor force has stalled and, recently, declined. This has serious implications for many middle-income families, as adding a second earner dramatically increases the financial resources available to a family. Falling work rates among women, alongside a decline in marriage and a dramatic rise in children raised by only one parent means that many families must get by with lower earnings:

The sharp decline in marriage and rise in single parenthood is affecting everyone except those with college degrees. The majority of first births to women without such degrees now occur outside of marriage. Given the importance of a second earner to family incomes, the rise in single parenthood and the recent decline in female labor force participation pose serious threats to continued middle-class prosperity.

conclusion

These statistics are sobering, and they imply that there is a great deal of work to be done to improve the prospects of the American middle class.

The challenge will be to improve those prospects in the face of changes in trade and technology that are destroying some jobs while creating others, yawning racial disparities, uneven economic growth from one place to another, and massive changes in gender roles and the family.

Over the course of the Future of the Middle Class initiative, we invite you to engage with us as we delve more deeply into these topics in a search for the best ways to create a stronger middle class.

      

On Mother’s Day, we celebrate all that mothers do for their children. Quite right, too. Being a mom requires plenty of sacrifice. Mothers also pay a price in terms of their own earnings. Here are some stylized facts about the “motherhood wage penalty:”

• It is real.
• There is no corresponding “fatherhood wage penalty.” If anything, men appear to receive an income bump in response to having a child.
• It reflects, in part, these facts: women are more likely to leave their jobs, reduce their hours, or take on more flexible or less-intensive (and thus lower paid) positions in response to having a child.
• But a large portion of the penalty is unexplained by job experience and seniority, implying employer discrimination or impacts of motherhood on productivity.

What is the “penalty” in the wage penalty for moms?

But there is a lot of variation in what researchers mean by the term “penalty.” It can be seen as:

1. The impact on a mother’s wages at a single point in time, relative to a similar woman who did not have a child. Defined this way, the penalty is about 5 percent of wages per child.
2. The absolute loss of lifetime earnings as a result of having a child.
3. The impact of parenthood on the gender wage gap – here, the motherhood penalty refers to the relative impact of having a child on women’s versus men’s wages
4. The way parenthood deferentially affects spouses’ wages; in other words, the gender wage gap within specific heterosexual couples.

How to minimize the motherhood penalty? It depends.

Understanding the penalty through each of these lenses can help inform policy and individual decisions about when and under what circumstances one should have a child. However, it is important to be clear about which “penalty” a study is estimating.

One question that many young women may ask, for example, is: what is the optimal age for becoming a new mother, if you want to minimize the impact of having a child on your earnings? The answer from the literature is…it depends. Are you trying to maximize your lifetime earnings, or close the pay gap with your spouse? Depending on your goals, the research might lead you to vastly different conclusions.

One recent Census Bureau paper made headlines for its finding that women who have their first child either before 25 or after 35 are much more likely to eventually close the pay gap with their spouse. That is to say, having a child during one’s prime career-building years (25-35) is likely to permanently exacerbate the pay gap between husband and wife:

^{Source: Chung et al. (2017)}

So, if you’re concerned about reducing wage inequality within your (heterosexual) marriage, it’s better to have a child very early or a bit later in your career.

But, perhaps you are less concerned about your income relative to your spouse than maximizing your lifetime earnings in absolute terms. In this case, you should avoid motherhood in your early 20s, especially if you are highly skilled. For women scoring in the top third of AFQT scores, waiting to have children until their 30s could reduce the lifetime motherhood penalty by more than one third, according to research by Elizabeth Ty Wilde, Lily Batchelder and David T. Ellwood. This is because the earnings hit resulting from motherhood is particularly consequential early in one’s career, when earnings are usually rising faster.

Other studies draw similar conclusions: young mothers experience a larger earnings hit (relative to not having children) than those who wait to have children. So, if you want to reduce the impact of having children on your lifetime earnings, it might be best to wait. On the other hand, if your primary concern is wage equity with your spouse and you simply can’t wait until your mid-30s to become a mother, you might choose to go ahead and have a child before you’ve really established yourself in your career. Of course, timing pregnancies precisely is sometimes difficult, but the use of the most effective forms of contraception, especially LARCs, makes it much more possible.

Reducing the penalty

Regardless of how these studies define “penalty,” the major conclusion is overwhelming: motherhood will cost you. For women, having a child interferes with the accumulation of human capital, and thus leads to lower incomes. This reflects the strongly gendered division of labor when it comes to child rearing. Mothers might cut back on working hours or education, choose more flexible (and lower-paid) positions, or drop out of the labor force entirely to confront their parenting responsibilities. At the same time, employers may discriminate against working mothers, who are perceived as less competent or committed to their jobs than men or women without children, and are thus penalized in the labor market through lower pay and less likelihood for advancement.

Is there a role for policy here? Given the extremely high costs associated with raising children, and the fact that women bear the brunt of childcare responsibilities, subsidizing the costs of childcare through the tax system could help. So too could paid parental leave which should encourage parents, and mothers in particular, to remain attached to the labor force while they take time away from work to care for new children. Cultural explanations related to gender norms and employer discrimination will be trickier to tackle, but family-friendly policies such as these can help nudge norms.

Mothers make huge sacrifices for their children. Perhaps the best thanks we can give them this Mother’s Day is to work towards reducing the sacrifices they make in terms of wages, too.

      

Growing gaps in family structure, educational investments, school readiness, test scores, and college entry and completion all make upward economic mobility a more difficult prospect for children born to poor families. Poor children in poor neighborhoods are at an even greater disadvantage. Growing up in an impoverished community doesn’t only affect your lifetime earnings – it can also affect the length of your life. It can even affect the quality of the air you breathe and the water you drink.

Limited income mobility among poor and minority children is often linked to differences in social environment, be it family, peer, school, or neighborhood environment. But income and race are also closely associated to the quality of one’s natural environment. Disparities in exposure to environmental toxins and pollution are more frequently the subject of the natural sciences, but their deep relationship with socioeconomic factors demands greater attention from social scientists. Indeed, a deep body of scientific research shows a strong and persistent relationship between socioeconomic status and exposure to environmental hazards. There is a good case for measuring poverty across multiple dimensions. Perhaps it’s time to start thinking about environmental poverty as an additional dimension of poverty.

Low-income and minority households are more likely to live in neighborhoods exposed to higher levels of water and air pollution.

poverty and pollution

The quality of America’s natural environment has greatly improved since the Clean Air Act was expanded and the Environmental Protection Agency established in 1970. Later legislation, like the Clean Water Act and the Safe Drinking Water Act, established quality standards for our nation’s water resources. Thanks to the enforcement of these laws and other efforts to improve or maintain environmental quality, Americans can worry less (for the most part) that their drinking water might catch fire or that the air they breathe contains dangerous levels of particulate matter (though recent rollbacks in environmental regulations might reverse these trends).

As the residents of Flint, Michigan learned, these risks have not disappeared. They are not the only ones exposed to dangerous levels of pollution or toxins. In 2015, nearly one-quarter of Americans were served by water systems that violated some component of the Safe Drinking Water Act.

Low-income and minority households are more likely to live in neighborhoods exposed to higher levels of water and air pollution.

No surprise, levels of concern over drinking water quality are higher among low-income and minority Americans.

Race and calss gaps in air quality

The ongoing Flint water crisis has drawn national attention to the dismal quality of drinking water systems in certain communities, but similar disparities exist regarding air quality. Low-income communities of color are likely to be exposed to far more pollution:

A 2008 study by Liam Downey and Brian Hawkins showed that lower-income households tend to live in neighborhoods with worse air quality, but the study’s major finding was that, even across income groups, black households are more likely to live in neighborhoods with poor air quality than white households. A more recent analysis demonstrated that toxic air pollution in urban areas was correlated with segregation. That is, the more segregated a neighborhood, the worse its air quality.

toxic for opportunity

These environmental inequalities have social and economic implications. Pollution can have substantial impacts on human health and, relatedly, the economy. Air pollution is a contributor to asthma and other respiratory diseases, which contribute to increased hospital admissions and lost working days. Water pollution increases the risk of bacterial, neurological, and viral infections and can contribute to higher rates of infant mortality. Researchers at the University of Minnesota determined that the average exposure to nitrogen dioxide (NO₂), a common outdoor air pollutant, is 38 percent higher among minorities nationwide, and bringing this exposure level down to that experienced by whites would be akin to reducing heart disease mortality by around 7,000 deaths per year. These health impacts can have economic consequences. For example, the social costs of air pollution (which includes the health costs of morbidity and mortality) resulting from domestic energy production alone amounted to $131 billion in 2011 – and that’s a substantial decrease from a decade prior.

There is a tricky game of causation or correlation here (do industrial plants move to low-income neighborhoods because property is cheap, or does their presence lower property values and attract low-income residents?) – but the association is clear and important. Concentrated inequality, segregation, and poverty disproportionately expose residents to environmental pollutants, adding to the list of disadvantages faced by poor and minority Americans.

Environmental health reflects and reinforces economic and racial inequality. There is a deepening understanding of the “social determinants of health,” thanks to the work of scholars like our Brookings colleagues Ross Hammond, Dayna Bowen Matthew, and Stuart Butler. But there are environmental determinants, too, and not only of health but potentially of opportunity. Earth Day is about the environment, of course – but it is also about equity.

      

It is hard to overstate the importance of the new study on intergenerational racial disparities by Raj Chetty and his colleagues at the Equality of Opportunity Project. Simply put, it will change the way we think the world works.

Making good use of big data—de-identified longitudinal data from the U.S. Census and the IRS covering nearly the entire U.S. population from 1989 to 2015—the Chetty team finds that while the economic differences between whites and Asians and whites and Hispanics are converging over time, black children have earnings disparities that persist across generations. These gaps are explained almost entirely by large gaps in employment rates and wages between black and white men, even after controlling for parental income.

Place matters

But the most surprising findings on the intergenerational gap between black and white men are found in their data on neighborhood effects.  Chetty’s previous research clearly demonstrates that rates of intergenerational mobility hinge on where children grow up. So a focus on the neighborhood a child grows up in is justified.  Therefore, in this new study, Chetty and his colleagues begin with the assumption that within metro areas, black and white children grow up in very different neighborhoods, which could account for their different outcomes.  To test this hypothesis, they compare the outcomes of children living in the same neighborhoods. They find that “the vast majority of the black-white gap persists even among boys growing up in families with comparable incomes in the same neighborhood; differences in neighborhood quality explain at most 25% of the black-white gap.”

But we should take a pause here. Most social scientists working in this field would say that compelling results showing neighborhoods accounting for up to a quarter of the black-white gap are in fact pretty important. While not one of their headline findings, Chetty and his colleagues show that neighborhoods have causal effects on racial disparities.

Measuring barriers in black neighborhoods

The Chetty team has a range of proxies for neighborhood quality. But even with their rich datasets, they may not have been able to fully capture the structural conditions of black neighborhoods. As Robert Sampson and I argued in 1995, it is difficult or in some cases impossible to reproduce in white communities the structural circumstances under which many black Americans live, including the historical legacy of extended racial discrimination and segregation across generations.

Consider what cities like Chicago, Baltimore, Milwaukee, Philadelphia, Pittsburgh, St. Louis, Cleveland, and Detroit have in common: they include many poor black neighborhoods that have undergone significant depopulation since 1970. The most visible symbols of this depopulation are abandoned buildings and vacant lots. The fundamental cause of this depopulation was the end of the Great Migration—that is, the cessation of black migration from the South in the early 1970s—and, simultaneously, the gradual movement of higher-income blacks from inner-city neighborhoods to other parts of the metro area, which I discussed in my book The Truly Disadvantaged (1987/2012).  The ranks of those who abandoned these neighborhoods were no longer being replaced by new black migrants from the South, nor were other ethnic and racial groups moving into many of these neighborhoods—hence the depopulation.

This urban depopulation makes it more difficult to sustain basic institutions or to achieve adequate levels of neighborhood social organization, factors that are also related to greater joblessness and higher crime rates. Very few urban white neighborhoods, even those with the same poverty rates as black neighborhoods, approximate these conditions.

Surprising to many, Chetty and his colleagues also find that the usual suspects such as education, marital status, and wealth explain very little of the intergenerational racial gap in income. They also infer that “ability” is not a factor, since while there is no significant gap in black male and black female test scores, there is a large gap in individual upward mobility rates.

Black fathers matter—a lot, but in a surprising way

What explains the limited upward mobility of black boys from certain neighborhoods? Perhaps the most striking finding of the whole report is the impact of “father presence” in census tracts on the mobility chances of black boys. Note that the researchers are not showing here the direct effect of a boy’s own father, or the marital status of his parents. This is about the broader presence of fathers in a given neighborhood. Note, too, that the finding relates specifically to fathers, not just men in general.

This is a wholly new and important finding. It is a highly specific, race-by-gender effect that provides direction for further empirical research and theoretical development on the causal mechanisms underlying these findings. Future work should include the investigation of whether the presence of black fathers in a census tract:

• reduces disciplinary problems;
• provides supervision, as well as direct mentoring support;
• generates more pro-social behavior and norms that lead to positive aspirations; and
• leads to changes in the way black boys are perceived and treated by peers.

It would also be useful to explore the role of father presence in the black inner-city neighborhoods of Chicago, Baltimore, Milwaukee, St. Louis, Philadelphia, Pittsburgh, Cleveland, and Detroit. These tend to be depopulated, as I previously discussed, in comparison with the inner-city black neighborhoods in places like New York, Boston, Los Angeles, and Washington, D.C. that tend to be more densely populated, not depopulated.  My hypothesis is that the more densely populated neighborhoods in these latter cities will feature a greater number of fathers, partly due to the presence of a higher proportion of working married-couple families.

Black and white children in vastly different neighborhoods, still

The paper also shows that the black-white mobility gap tends to be relatively small in neighborhoods with low poverty rates (below 10 percent) and high rates of fathers present (more than half). This demonstrates that there are childhood environments that yield good outcomes for black males. The problem is that fewer than five percent of black children grow up in such environments. This is a startling finding, given the growing economic inequality in the black community.

This inequality can be seen in differences within racial categories in the Gini coefficient, a measure of income inequality that ranges from 0 (perfect equality) to 1 (maximum inequality). Although the absolute level of black income is well below that of whites, blacks nonetheless display the most intra-group income inequality, with a household Gini index of 0.50 in 2016, followed by whites and Hispanics at 0.47, and Asians at 0.46.

Indeed, one of the most significant changes in the past several decades is the remarkable gains in income among more affluent blacks. The percentage of black households with incomes of at least $75,000 more than doubled from 1975 to 2016, to 24 percent, adjusting for inflation. Those making $100,000 or more quadrupled to 15 percent. In contrast, white households saw a less impressive increase, from 13 to 31 percent. On the other hand, the percentage of black households with incomes below $15,000 only declined by six percentage points, to 20.6 percent, between 1975 and 2016.

Income inequality is related to income residential segregation, and a 2014 study by the sociologists Kendra Bischoff and Sean Reardon shows that income segregation in metropolitan areas with populations of more than 500,000 has grown rapidly in the last several years, especially among black families. Indeed, whereas black American families in 1970 recorded the lowest rates of income residential segregation, they now register the highest income segregation of all major racial and ethnic groups. The rising income segregation in the black community is driven by both the growth of affluent blacks and the deteriorating conditions of poor blacks.

Race, class, gender, place: all connected

On the one hand, then, race trumps class when the focus is on interracial differences in neighborhoods and neighborhood effects, as the research of Chetty and his colleagues so clearly demonstrates. But on the other hand, class trumps race when the focus in on changing intra-racial neighborhood differences, especially those within the black community. In each instance, it is the interaction between race and class that researchers and policymakers have to address, not a focus on one to the exclusion of the other.

      

For more than half a century, the term “the American middle-class,” has served as a political reference to white American upward mobility. This was less an artifact of particular calculations than one of historical experiences and demographic realities. Since at least the 1950s, Americans who were neither wealthy nor “disadvantaged” were, by default, middle class. Who fell into this category? With most African-Americans and Latinos frequently restricted in their access to high quality education, jobs, or homes and because, over the last 50 years, the U.S. has had a majority white population, it could be assumed that the majority of middle-class Americans were white. So, the term “American middle class,” while not historically and intentionally located in a discourse about race, has always inherently been about race, specifically about white Americans. Fast forward to 2016 and two trends have emerged which highlight the degree to which the “American middle class” can no longer serve as an implicit proxy for a group that is predominantly white. First, the demography of the nation is changing rapidly. As our colleague Bill Frey has demonstrated, the U.S. is well on its way to becoming a country where black and brown diversity predominates. In 2017, for the first time, the majority of American children under 10 were black and/or brown. From a purely mathematical perspective, this implies that at some point in the not too distant future, the composition of the American middle class will begin to mirror that of American society as a whole. Second, in the context of greater economic inequality in the U.S., a recent recession, and a 2016 Presidential election outcome that highlighted the plight of low-income white Americans, there is a heightened general public awareness that some previously middle-class whites are no longer “middle class.” We are then at a political inflection point. Who is the middle class? What does the changing racial composition of the middle class mean for politics, or policy? Our goal here is to provide some empirical grounding to this debate. We describe the changing racial composition of a group we define as the middle class: those between the 20th and 60th percentiles of the income distribution. The middle class according to this definition was predominantly white in 1980, but is only just majority white (56 percent) today, and will be race-plural (what some label “majority minority,” a phrase we dislike) by 2042 (see Note 1). Within the next quarter century, we estimate that whites will no longer be the majority of the middle class, as Hispanics and other minority groups age into the adult population (see Note 2).

who are the middle class?

To say that there are many definitions of the “middle class” would be an understatement. Politicians, understandably enough, avoid being too specific in their use of the term: they want as many people as possible to believe they will benefit from a “tax cut for the middle class” or “raising the living standards of middle class” – that is, by ensuring that middle class families achieve common aspirations like home and car ownership, higher education, health and retirement security and time for leisure. This is not especially helpful, given that most Americans, of all backgrounds, share these aspirations. Most Americans define themselves as middle class, although the number varies significantly depending on how the survey question is constructed. Most polling organizations end up dividing the middle class into sub-categories, usually “lower middle class”, “middle middle class” and “upper middle class.” Even then, the “middle middle” is always the most popular choice. Scholars define the middle class in a number of different ways. Many emphasize cultural factors including values and aspirations; others highlight power relations. Researchers with an empirical bent might delineate class by absolute or relative income, or household wealth. For an excellent overview of these definitional questions, see Income Inequality: Economic Disparities and the Middle Class in Affluent Countries, edited by CUNY’s Janet Gornick and Markus Jantii. (We’ll have more to say on this definitional issue in a later piece). For the purposes of this analysis, we consider families whose “head of household” is prime-age (ages 25 to 54) as “middle class” if they fall within the second and third household income quintiles (i.e. between the 20th and 60th percentiles). This group broadly consists of people who are above the official poverty line, but far from prosperous – with incomes ranging between about $30,000 and $86,000. It is important to acknowledge that there is nothing magical about any particular definition. Each comes with advantages and disadvantages. One reason we focus here on this 40 percent of the population is that this group has seen the slowest income growth in recent decades: To repeat: we do not claim that our definition of the middle class is the definition. We should also be clear that, in part, our definition is influenced by the economic trends described above, rather than being constructed in splendid isolation.

The racial pluralism of the middle class

In 1980, non-Hispanic whites composed over three-quarters of the middle class – 23 percentage points more than they do today. Some of this change is simply due to the increasing size of racial and ethnic minorities across all income groups, some is due to rising incomes, and some due to differing age compositions: the median age of non-Hispanic whites is 43, compared to 29 among Hispanics. The white share of the middle class fell to 55 percent in 2017, and will drop below half to 49 percent within the next quarter-century, according to our calculations: A note on the projection. We have assumed that the relationship between relative income status and race remains unchanged between 2017 and 2042. The increased racial diversity of the middle class in 2042 is, then, the mechanical result of a broader demographic shift. Many fewer Americans aged between 25 and 54 in 2042 will be white, and many more will be other races and ethnicities, especially Hispanic (see Note 3). These specific estimates should be taken with a few large grains of salt – we did not account for divergent income trajectories, possible changes to immigration policy, or differing life expectancies by race. But predicting such changes would be a difficult undertaking, so we have adopted a simpler strategy.

identity politics versus the middle class: a false choice

There is an important political debate taking place over the plight of the so-called “forgotten Americans” (look out for an important new book from Isabel Sawhill on this topic later this year). Scholars and politicians on both left and right, and in bipartisan groups such as the AEI/Brookings Working Group on Poverty and Opportunity, have wrestled with the challenges facing middle- and lower-income households. Among Democrats in particular, there is a particularly robust argument over the balance between “identity politics” – focusing on the specific challenges faced by groups such as women, racial and ethnic minorities, or LGBTQ Americans – and the white middle or working class. As Columbia Professor Mark Lilla writes:

In recent years American liberalism has slipped into a kind of moral panic about racial, gender and sexual identity that has distorted liberalism’s message and prevented it from becoming a unifying force capable of governing.

On the other side of the debate, Matthew Yglesias of Vox writes (in part in response to Lilla) writes:

People have identities, and people are mobilized politically around those identities. There is no other way to do politics than to do identity politics.

These are important discussions. But we should be careful not to set up arbitrary distinctions. The middle class, at least defined in economic terms, is not white: it is racially diverse; and whites will make up less than half the middle class within a few decades. Any policies that help the middle class will therefore help black and Hispanic Americans as much as whites. The choice between identity-based and class-based politics is, to this extent, a false one.

Notes

1. The term “majority minority,” while common, is unhelpful. First, when there is no single racial group in the majority, there is by definition no majority. Second, the phrase lumps together all minorities who are not white into a single category (non-whites), which is inappropriate given the significant differences between these groups. Hence our preference for the term race-plural.
2. Note that for the purposes of this paper we follow common practice in referring to Hispanic as a racial rather than ethnic category.
3. We look first at the number of individuals within each race/ethnic category under the age of 29 in 2017. We then multiplied these values by the percentage of each race/ethnic category who fell into the prime-age middle class as of 2017. For example, 36 percent of prime-age non-Hispanic whites were middle class in 2017. We multiplied the number of whites ages 0-29 (about 66.2 million) by 36 percent to estimate the number of prime-age whites who would be middle class in 25 years. We then divided this number by the total size of the prime-age middle class (126.2 million) to obtain our final figure of 49 percent.

      

The proportion of U.S. adults in paid work has declined in recent decades. While the fall in male employment gets the most attention, female work rates are declining too. A new NBER paper from Katharine Abraham and Melissa Kearney provides a comprehensive review and rigorous analysis of the overall trends, and potential contributory factors including trade, technology, weak demand, and an aging population to name a few

Abraham and Kearney suggest that other factors are likely playing a role, too, which are impossible to quantify. These include changes in the social norms around paid work and “improvements in leisure technology” (read: Rainbow Six Siege and other awesome games). Note that their preferred measure of labor market activity is the employment-to-population ratio (see Eleanor Krause and Isabel Sawhill’s guide for a quick explanation of the different employment metrics).

The Abraham and Kearney paper is a gold mine, and an important contribution to a vital debate. But one sentence in particular jumped out to us: “The decline for men age 25 to 34 (5.6 percentage points) was more than twice as large as the decline for women the same age (2.3 percentage points); among those age 35 to 44 and those age 45 to 54, the declines for men and women were more similar.”

For all the worries about middle-aged men, it is actually men at the younger end of the prime-age years who have seen the sharpest drop in employment rates:

These are not older men who have been dislocated from traditional manufacturing jobs after years of service. They are younger men who seem to be struggling to connect to the labor market at all. Perhaps these men are poorly educated? We know, after all, that women have overtaken men in educational attainment. While both men and women aged 25-34 in 2016 were more educated than in 1999, the educational improvement has been faster for women:

The negative impact of low levels of education on work rates is significant – and essentially the same for men and women. The employment rate among young adult women with just a high school diploma has dropped by 8.9 percentage points. For men of the same age group, the fall is 9.6 percentage points. Since there are more men with less education, this explains some of the gender difference in employment rate changes.

But it is actually among the better educated that the gender gap emerges. Among those aged 25-34 with a college degree, the male employment rate has dropped twice that of women:

Perhaps these men are more likely to be impacted by some of the factors discussed by Abraham and Kearney. Perhaps their specific educational credentials did not fit well with the needs of the labor market. Perhaps more of them are stay-at-home dads. We do know that at least some of them are living with their parents, and so may be under less pressure to find paid work. As Aguiar et al. (2017) report, two out of three non-employed young men ages 21 to 30 lived with a parent or close relative in 2015, compared to fewer than half in 2000. Perhaps the relative attraction of the labor market compared to video gaming is less for men than for women.

Our interest in this particular, small slice of the population – young men with some college education – should not distract from the broader findings of the paper: that large economic forces, especially trade, technology, and broad demographic shifts including an aging population are behind much of the fall in employment. For both men and women, a post-secondary education is increasingly a requirement in the modern labor market.

But the sizable drop in employment rates among younger men (25 to 34) should also give us pause, perhaps especially the drop among those who have been to college, but have not successfully made the transition from learning to earning.

      

There is not very much to like about the Tax Cuts and Jobs Act of 2017. It delivers big benefits to the affluent, creates new loopholes and complexities, and will send the deficit soaring. One provision with some merit, however, is the introduction of a tax on the endowments of wealthy colleges.

Of course, it has hardly gone down well within the Ivy League. But there are some hard truths to face here. A small handful of elite universities have aggressively raised billions of dollars supported by tax-free endowments while increasing tuition costs and doing little to promote social mobility. The endowment tax should not simply be used to raid their funds. It should be used as a tool to incentivize colleges to become more inclusive. Less hoarding of both money and opportunity; more investments in upward mobility. If structured properly, the threat of taxation could be a force to create more opportunity for children from working families, not an offset to cut taxes for children of multi-millionaires.

the harvard hedge fund

Total student tuition paid to Harvard was just under $300 million, less than one percent of the size of their endowment. Harvard and other elite schools endowments pay more in management fees than they pay out in scholarships. Most people think of Harvard University as an educational institution. But from a distance it looks worryingly like a large hedge fund, operating a small university on the side in order to secure non-profit status. Harvard’s endowment is over $37 billion, more than 7 times as large as the university’s operating budget, and more than half the total spending each year by America’s 1,462 community colleges. Harvard’s highest paid employee, earning almost $15 million last year, is the manager of the endowment. (The top seven employees managing the endowment earned a total of $58 million between them).

The Big Four in terms of endowment size – Harvard, Yale, Stanford and Princeton – have fewer than one in five students receiving Pell Grant support.

This is not just an Ivy-covered issue. A similar logic could be applied to ask if other schools are simply professional football teams with larger non-profit arms. For example, the University of Alabama’s head football coach is paid more than $11 million a year; the University of Michigan’s received $7 million in 2017.

Of course, these institutions are not operating in a vacuum. They face a range of perverse incentives and market pressures, including college rankings that promote endowment growth at the cost of student aid.  Endowment-to-student ratios are a common factor used in college rankings, resulting in “my endowment’s bigger than your endowment” arms race. The government student lending program provides little if any counter-balance.

The new small tax on endowments will not solve these problems. But it does serve as a wake-up call to higher education leaders: time to return to their core values of providing opportunity and enrichment for students. Properly applied, the tax could help to promote these values. Many colleges could do much more to attract and retain lower-income students with little impact on their endowment.

use an endowment tax as a stick – and a carrot

There is a strong case for public policies that encourage funds to be used on students, and especially on students from poorer backgrounds, rather than growing massive tax-free endowment funds. The Republicans, however, have not used tax reform to alter behavior or change incentives. They have simply gone for the money; with the political side-benefit of tweaking the noses of the learned liberals dominating the top ranks of elite college. To the GOP, college endowments are an ATM, with the money taken out given straight to corporations and the rich.

A better approach would be to use an endowment tax as a means for altering incentives in a pro-mobility direction. Rather than punishing all colleges with large endowments equally, the goal should be to reward those which are using their funds in an opportunity-enhancing way. The colleges who will face the new tax vary greatly in terms of the composition of their student bodies:

At one end of the spectrum is Berea College, a school that uses its endowment to provide free education for a diverse student body: two out of the three are the first in their family to attend college and 83 percent are Pell Grant recipients. Meanwhile the Big Four in terms of endowment size – Harvard, Yale, Stanford and Princeton – have fewer than one in five students receiving Pell Grant support.

In general, selective and elite colleges serve to reinforce existing social and economic inequalities, rather equalizing opportunities. The postsecondary education system as a whole plays an important role in the reproduction of inequality across generations. There is more dream hoarding than dream sharing going in most of our top colleges, whether they like to admit it or not.

We don’t propose here to set out a concrete policy proposal. But here are some possible avenues:

1. Provide an exemption from or reduction in the endowment tax for colleges and universities who are working hard to promote opportunity. One option is to use a certain threshold for Pell-eligible students (say at least 33 percent); another would be to base a threshold on the Mobility Scorecards produced by the Equality of Opportunity Project.
2. Keep the endowment tax, but dedicate all revenues to proven pro-mobility policies in post-secondary education, including scholarships, tutoring schemes or programs like CUNY’s ASAP, just successfully replicated in other sites.
3. Exempt endowment gifts that are for explicitly inclusive purposes. A considerable chunk of the endowment funds are restricted to certain kinds of spending – usually at the request of a donor. Exempting pro-mobility endowment funding (eg. Dedicated to scholarships for low-income students or similar) from the tax could encourage a shift in the norms and behavior of donors.
4. A reduction in the endowment tax could be offered to institutions who abandon unfair and anti-meritocratic practices such as admissions preferences for legacies or children of donors.

As we said, these are just food for thought. Our hope is that colleges impacted by the new tax will not simply rail against its injustice, or the evils of the Trump administration in general. That would play into Republican hands. Note that most Republican or Republican-leaning voters now think that colleges “have a negative effect on the way things are going in the country,” according to a 2017 Pew poll.

The leaders of our richest colleges and universities must do more than simply hunker down. There are valid criticisms of the way they are amassing and deploying their financial resources. A tougher version of the endowment tax from House Republicans was beaten back. But this isn’t over. Colleges would be wise to act now.

      

This is the third blog post for the 2018 series on paid family leave jointly sponsored by AEI and Brookings. Aparna Mathur at AEI and Isabel Sawhill at the Brookings Institution are the co-directors of the AEI-Brookings Project on Paid Family Leave. The project includes a diverse group of individuals from different organizations with expertise on this topic. Following our initial blog to tee off the series, we have invited each of the working group members to offer their thoughts on the topic.

We invite you to engage with us as the administration works to chart out policies on these topics. This will help us better inform policymakers of the practical day-to-day realities of living in a country where millions lack access to paid leave at the birth of a child or to meet other caregiving needs.

Beginning on October 1st, 2018, workers in Austin, Texas will have a new right: to paid leave for sick days, either for themselves or a close family member. Austin is the 31st city to adopt a paid sick leave policy – but the first in the South. (Nine states and two counties have also passed paid sick leave laws). The new provision, City Code Chapter 4-19, will allow workers to claim one hour of sick leave for every 30 hours worked, up to an annual cap of 64 hours for companies with more than 15 employees, and 48 hours for those with 15 or less.

There is a strong case, as made by my colleagues Eleanor Krause and Isabel Sawhill, for a federal sick leave policy. But in the current political climate, federal action seems unlikely, and in the meantime, a patchwork of different laws is growing up across the country.

Sick leave and family leave laws vary between different jurisdictions in a number of different ways – including in the way that they define “family.” Grandparents qualify for paid family leave in the State of New York; but not across the river in New Jersey. In Washington State, you can take time off to care for a sick sibling; but not in California. (See this very useful guide to some of the state paid family leave laws, from the National Partnership for Women and Families).

The definition used in the new Austin provision for sick leave is very broad. Section 4-19-1(e) reads: “FAMILY MEMBER means an employee’s spouse, child, parent, or any other individual related by blood or whose close association with the employee is the equivalent of a family relationship.”

The “any other individual related by blood” category is potentially very large, depending on how narrowly the term “related” is to be used. Including those “whose close association with the employee is the equivalent of a family relationship” widens the pool even further. At this point, the definition has also become circular, with a “family member” defined as a person with whom the employee’s relationship is “the equivalent of a family relationship.”

At the federal level, in the Healthy Families Act (H.R. 1516) sponsored by Rep. DeLauro (D-CT) would provide a right to paid sick leave, including to care for a sick “domestic partner.” But what does term “domestic partner” mean? According to H.R. 1516, it means “another individual with whom the individual is in a committed relationship.”

So: what counts as a “committed relationship”? H.R. 1516 provides an answer:

“A relationship between two individuals, each at least 18 years of age, in which each individual is the other individual’s sole domestic partner and both individuals share responsibility for a significant measure of each other’s common welfare. The term includes any such relationship between two individuals, including individuals of the same sex, that is granted legal recognition by a State or political subdivision of a State as a marriage or analogous relationship, including a civil union or domestic partnership.”

In a later section of the Bill (Section 5 b 3), acceptable use of the paid sick leave includes “an absence for the purpose of caring for a child, a parent, a spouse, a domestic partner, or any other individual related by blood or affinity whose close association with the employee is the equivalent of a family relationship.” (Those are my italics). As in the Austin law, the definition has circled back on itself.

The tortuous language makes many of these laws virtually impossible to enforce in terms of definitions. Should an employer wish to, they can check if somebody is married or has a child. The difference between a domestic partner, boyfriend or roommate will be difficult to prove.

The attempts in H.R. 1516 and the new Austin law to pin down an acceptable definition are exercises in futility. What if two people “share responsibility for a significant measure of each other’s common welfare” but prefer to live apart? What if two people end up living together out of economic necessity, even though their relationship is over? Who says only a “sole” person is allowed to play this role in my life?

The instinct behind many of these provisions is an inclusive one, especially with regard to same-sex couples. Some of the state plans pre-date the 2015 decision by the Supreme Court obliging all states to recognize same-sex marriage. But now that same-sex marriage is legal, these expansive definitions are arguably less necessary. That was, after all, part of the point of the campaign – to help secure the same legal rights for same-sex couples.

Right now, the attempts to define a family member in paid sick and paid family leave laws fall between two stools. They are too broad to be sensibly enforced, but still prescriptive enough to exclude some worthy cases. Why should I not be able to use some of my sick leave to care for my oldest, dearest friend from my church?

Lawmakers should either narrow their definition of family member so that it would have some chance of being enforceable in a court; or leave it to individuals to decide.

The narrowing option would mean restricting eligibility, for example, to spouses, rather than domestic partners or close associations or civil unions, given that lesbians and gay men can now choose to marry.

Better still, lawmakers should just get out of the business of attempting to delineate acceptable family relationships, and leave it the individual employee. Then any law could simply state that the sick leave could be taken by the worker for either their own sickness or “to care for another.”

If the government is to grant workers a right to time off to care for someone who is sick, perhaps it should also grant them the right to choose who most needs that care.

      

“Big data” is particularly useful for demonstrating variation across large groups. Using administrative tax data, for example, Stanford economist Raj Chetty and his colleagues have shown big differences in upward mobility rates by geography, by the economic background of students at different colleges, by the earnings of students taught by different teachers, and so on.

The basic message here is that when it comes to understanding how to promote more opportunity, it’s dangerous to make generalizations. For any given national trend or picture, there will be places, people, or institutions that do much worse than the average—and plenty that do a lot better.

A few weeks ago, Raj Chetty stopped by Brookings to present on his most recent research. After his presentation, he joined me and Adrianna Pita on the Intersections podcast to discuss how big data helps us understand diversity within populations—or as academics would say: helps us demonstrate heterogeneity.

You can listen to the podcast here. If you missed Chetty’s presentation, you can watch video from the event.

Chetty’s not the only one working with large datasets, of course. Stanford’s Sean Reardon has assembled huge datasets of test score results, at different points in time, for schoolchildren across the nation. Armed with 300 million test scores across more than 11,000 school districts, Reardon looks, for example, at results in Grade 3 and Grade 8. As you would expect, the school districts with lower scores at Grade 3 are more likely to be lower-income areas, and also to have lower scores at Grade 8.

More interesting in terms of policy, however, are the departures from this story. The chart below shows where the 100 largest school districts place in terms of achievement in Grade 3 and achievement growth between Grades 3 and 8. (The size of each bubble corresponds to the district’s population). Districts in the upper-right quadrant, like Plano Independent School District, have high initial achievement as well as high growth; districts in the upper-left, like Chicago, have low initial achievement but high growth; and so on.

Examining the data by district allows us to see that while third-graders in Baltimore and Chicago start with similar achievement levels, by eighth grade, the Chicago students have jumped ahead. Chicago eighth-graders score at about the same level as those in New York and Henrico County, despite starting from a lower base. (Districts that fall on the same diagonal line have the same average eighth-grade achievement level).

“The socioeconomic profile of a district is a powerful predictor of the average test score performance of students in that district,” says Reardon. “Nonetheless, poverty is not destiny: There are districts with similarly low-income student populations where academic performance is higher than others. We can – and should – learn from such places to guide community and school improvement efforts in other communities.”

The work of big-data scholars like Reardon and Chetty helps to illuminate the scale of the challenge of improving upward mobility for low-income kids. But the heterogeneity also gives hope, too. We can identify institutions or cities that are bucking the trend in a positive direction: and see what we can learn from them. Big data then becomes a tool not only for describing the world, but changing it.

      

The retreat from work among men is a topic of great concern for scholars and policymakers. And for good reason: over the past 50 years, the prime-age male employment rate declined by 10 percentage points. While men’s employment rates have dropped in many countries, a drop on this scale is unique to the U.S.

But the story on women’s work is just as important – indeed, arguably more so. In recent years, there has been a reduction in paid work among U.S. women, one that is not paralleled in other advanced nations. Getting more women into the labor market is perhaps an even higher priority than raising male employment rates.

The three main motivations for increasing employment rates are to close skill gaps, boost growth, and cut poverty. In each case, there are good reasons to focus on women’s work.

1. More WORKING women would narrow the skills gap

In an economy nearing “full employment,” many employers are increasingly concerned that workers do not have the right skills for today’s occupational demands. Getting out-of-work adults into the labor force could narrow this “skills gap.” But this is truer for women than for men. Education levels among women who are not employed are higher than for men. Half of the women who are not in paid work have at least some higher education:

The growth in middle-class jobs is also in occupations where women are more concentrated – those in the so-called HEAL professions (health, education, administration, and literacy). The fastest-growing occupations also require more advanced credentials than in the past. So employer complaints of skill shortages might be quelled if some of these well-educated women joined the workforce. Women may also have more of the so-called “soft skills” that employers say are in short supply, and fewer problems like drug addiction.

2. More working women would boost growth

All else equal, adding more workers to the economy will increase aggregate economic output. Bolstering the nation’s GDP is an important goal for policymakers of all political stripes. Improving work rates is one promising way to achieve this growth. Estimating the contribution of higher work rates to overall economic activity is necessarily difficult; it will depend on a range of factors, not least labor productivity.

Say that the employment rate of either men or women could be lifted by 10 percentage points. What would be the economic impact? It would depend of course on the kinds of jobs they took, and the wages they earned. Here we present some very rough estimates, under two different scenarios. In the first, we assume that each additional worker is paid the current median wage for their educational level and gender. In the second, we assume that women earn the same as men, conditional on education: in other words, we assume that there is no gender pay gap. In the first scenario, the economic impact of getting more men into work is greater, despite their lower average education, because they earn more. Under the second scenario (pay parity), the economic impact of getting more women into work is greater, because of their higher levels of education:

Under current conditions (i.e. with a gender pay gap), the economic impact of raising male employment rates is greater (to the tune of $46 billion) than increasing female employment by the same amount. Under conditions of pay equality, the opposite is true.

Under conditions of unequal pay, there is a then a stronger economic argument for raising male employment rates. But there is a circularity to this argument. When women work less, or take more time off work, say to care for children, they end up earning less. This dampens the economic impact of increased female employment; which could in turn be used as a rationale for prioritizing men’s work. And so the cycle turns.

3. More working women would cut child poverty

The best route out of poverty is a job. Wages and salaries make up more than three-quarters of U.S. household income, and much higher fractions among younger, working-age adults. Increasing work rates among low-income households is one of the surest ways to cut poverty. Policymakers are especially, and correctly, concerned about rates of poverty in households with children. From a child poverty perspective, getting more women into work matters more than boosting male work rates. This is because children live with women.

The employment rate of prime-age women with children under 18 in the household is 22 percentage points less than that of men (68 percent versus 90 percent). Some mothers choose to stay at home while their husbands work, but many do not live with a male breadwinner. The employment rates of unmarried mothers is 9 percentage points lower than for unmarried fathers. There are also many more families headed by out-of-work mothers than out-of-work fathers. Among families in which no parent works, out-of-work mothers outnumber out-of-work fathers nearly two-to-one:

Of course these patterns could change, just as the gender pay gap could narrow. In the future there may be as many single jobless fathers as single jobless mothers. But as things stand, it is women’s employment that will make the biggest dent in child poverty rates.

what can we Do about these trends?

None of this is to say that declining male work rates are not a problem. But it is important to be clear what the specific problem is. There may be other reasons to worry about male worklessness, above and beyond the economic ones discussed here. Some commentators fear that being without paid work is more damaging to the psyche, health, and status of men than of women. This may be true. In which case the goal is as much about affecting culture as economics. But if that is the real motivation behind efforts to increase male employment, it would be good to be honest about it.

Meanwhile, in terms of achieving more bread-and-butter goals like closing skill gaps or reducing child poverty, women seem to be a better bet than men. Women also seem to be more responsive to policy interventions aimed at increasing paid work (see, for instance, the recent evaluation by MDRC of Paycheck Plus).

There are other policy changes that might increase female employment, given that women continue to have more responsibility for raising children. The lack of paid family leave and adequate childcare supports are unique to the U.S., and can explain a substantial portion of the decline in female participation rates compared to other advanced economies. Low wages might also play a role – over 90 percent of “homemakers” who are not working (the vast majority of whom are female) contend that they would be willing and able to work if it paid just 10 percent more than their previous job.

Of course, growing the labor force isn’t a zero-sum game. Many of the policies that have been promoted to boost male work rates would benefit female work rates as well. Improved skills training, better adjustment assistance, higher wages, and greater workplace flexibility could all play a role.

Too often, discussions about employment proceed on the unchallenged basis that male employment rates matter more than female ones. But it is far from obvious why that should be the case. Getting more women into work is just as important – and perhaps more so.