America’s opportunity gap

This article appears in the November-December 2012 issue of Foreign Affairs magazine. Reprinted here by permission.

It’s Hard to Make It in America: How the United States Stopped Being the Land of Opportunity

by Lane Kenworthy

For all the differences between Democrats and Republicans that were laid bare during the 2012 U.S. presidential campaign, the parties’ standard-bearers, Barack Obama and Mitt Romney, do seem to have agreed on one thing: the importance of equal opportunity. In remarks in Chicago in August, Obama called for an “America where no matter who you are, no matter what you look like, no matter where you come from, no matter what your last name is, no matter who you love, you can make it here if you try.” The same month, he urged the Supreme Court to uphold affirmative action in public universities, putting his weight behind what has been a mainstay of U.S. equal opportunity legislation since the 1960s. Days later, the Republican vice presidential nominee, Paul Ryan, echoed Obama’s sentiment, saying, “We promise equal opportunity, not equal outcomes.” Romney, too, argued that whereas Obama “wants to turn America into a European-style entitlement society,” his administration would “ensure that we remain a free and prosperous land of opportunity.”

It is no accident that both campaigns chose to emphasize equality of opportunity. It has long been at the center of the American ethos. And one of the United States’ major successes in the last half century has been its progress toward ensuring that its citizens get roughly the same basic chances in life, regardless of gender or race. Today, women are more likely to graduate from college than men and are catching up in employment and earnings, too. The gap between whites and nonwhites has narrowed as well, albeit less dramatically.

Yet this achievement has been double edged. As gender and race have become less significant barriers to advancement, family background, an obstacle considered more relevant in earlier eras, has reemerged. Today, people who were born worse off tend to have fewer opportunities in life.

Of course, there is no perfect way to measure opportunities. The best method devised thus far is to look at outcomes: college completion, gainful employment, and sufficient income. If the average outcome for one group far outpaces that for another, social scientists conclude that the first group had greater opportunities. Comparing outcomes is not foolproof, as differences in outcomes can result from differences in effort. But a person’s effort is itself shaped by the circumstances he or she encounters.

To assess equality of opportunity among people from different family backgrounds, the measure of outcome that social scientists look at is relative intergenerational mobility — a person’s position on the income ladder relative to his or her parents’ position. Social scientists don’t have as much information as they would like about the extent of relative intergenerational mobility, its movement over time, and its causes. The data requirements are stiff; analysts need a survey that collects information about citizens’ incomes and other aspects of their life circumstances, then does the same for their children, and for their children’s children, and so on. The best assessment of this type in the United States, the Panel Study of Income Dynamics, has been around only since the late 1960s.

Even so, there is general consensus among social scientists on a few basic points. First, an American born into a family in the bottom fifth of incomes between the mid-1960s and the mid-1980s has roughly a 30 percent chance of reaching the middle fifth or higher in adulthood, whereas an American born into the top fifth has an 80 percent chance of ending up in the middle fifth or higher. (In a society with perfectly equal opportunity, every person would have the same chance — 20 percent — of landing on each of the five rungs of the income ladder and a 60 percent chance of landing on the middle rung or a higher one.) This discrepancy means that there is considerable inequality of opportunity among Americans from different family backgrounds.

Second, inequality of opportunity has increased in recent decades. The data do not permit airtight conclusions. Still, available compilations of test scores, years of schooling completed, occupations, and incomes of parents and their children strongly suggest that the opportunity gap, which was narrowing until the 1970s, is now widening.

Third, in a sharp reversal of historical trends, there is now less equality of opportunity in the United States than in most other wealthy democratic nations. Data exist for ten of the United States’ peer countries (rich long-standing democracies). The United States has less relative intergenerational mobility than eight of them; Australia, Canada, Denmark, Finland, Germany, Norway, Sweden, and the United Kingdom all do better. The United States is on par with France and Italy.

So how did the United States get here? Why did it falter where other nations have not? And how can it fix the problem? On the right, a standard proposal is to strengthen families. On the left, a recent favorite is to reduce income inequality. And everyone supports improving education. To know which proposals would work best, it helps to understand the roots of the new opportunity gap.


Between the mid-1800s and the 1970s, differences in opportunity based on family circumstances declined steadily. As the formerly farming-based U.S. labor force shifted to manufacturing, many Americans joined the paid labor force, allowing an increasing share of them to move onto and up the income ladder. Elementary education became universal, and secondary education expanded. Then, in the 1960s and 1970s, school desegregation, the outlawing of discrimination in college admissions and hiring, and the introduction of affirmative action programs helped open economic doors for an even wider swath of Americans.

But since the 1970s, the United States has been moving in the opposite direction. A host of economic and social shifts seem to have widened the opportunity gap between Americans from low-income families and those from high-income families. First, family life has changed, at least for some. The share of poorer children growing up with both biological parents has fallen sharply, whereas there has been less change among the wealthy. About 88 percent of children from high-income homes grow up with married parents. That is down from 96 percent four decades ago. Meanwhile, only 41 percent of poorer children grow up in homes with married parents, down from 77 percent four decades ago. That has hurt poorer children’s chances of success, since children who live with both of their parents are more likely, even accounting for income, to fare better in school, stay out of trouble with the law, maintain lasting relationships, and earn higher incomes as adults.

The modern culture of intensive parenting — a largely middle- and upper-class phenomenon — adds to the gap. Low-income parents are not able to spend as much on goods and services aimed at enriching their children, such as music lessons, travel, and summer camp. Low-income parents also tend to read less to their children and provide less help with schoolwork. They are less likely to set and enforce clear rules and routines for their children. And they are less likely to encourage their children to aspire to high achievement in school and at work.

Furthermore, a generation ago, most preschool-aged children stayed at home with their mothers. Now, many are enrolled in some sort of child care. But the quality of their experiences varies. Affluent parents can send their children to nationally recognized education-oriented preschools. Poorer parents might have little choice but to leave their children with a neighborhood babysitter who plops them in front of the television. Research by the economist James Heckman and others finds that much of the gap in cognitive and noncognitive skills between children from poor homes and those from affluent homes is already present by the time they enter kindergarten.

Things don’t improve once children reach grade school. Funding for public K-12 schools, which used to vary sharply across school districts, has become more even in recent decades. Nevertheless, a large difference remains in the quality of education between the best and the worst schools, and the poorest neighborhoods often have the weakest schools. According to data compiled by Sean Reardon of Stanford University’s School of Education, the gap in average test scores between elementary- and secondary-school children from high-income families and those from low-income families has risen steadily in recent decades. Among children born in 1970, those from high-income homes scored, on average, about three-quarters of a standard deviation higher on math and reading tests than those from low-income homes. Among children born in 2000, the gap has grown to one and a quarter standard deviations. That is much larger than the gap between white and black children.

Partly because they tend to be far behind at the end of high school, and partly because college has gotten so expensive, children from poor backgrounds are less likely than others to enter and complete college. The economists Martha Bailey and Susan Dynarski have compared the college completion rates of Americans who grew up in the 1960s and 1970s to the rates of those who grew up in the 1980s and 1990s. The share of young adults from high-income homes that got a four-year college degree rose from 36 percent in the first group to 54 percent in the second group. The share from low-income homes, however, stayed almost flat, rising only from five percent to nine percent.

When it comes time to get a job, the story is no better. Low-income parents tend to have fewer valuable connections to help their children find good jobs. Some people from poor homes are further hampered by a lack of English-language skills. Another disadvantage for the lower-income population is that in the 1970s and 1980s, the United States began incarcerating a lot more young men, including many for minor offenses. Having a criminal record makes it all the more difficult to get a stable job with decent pay — if, that is, good jobs still exist. A number of developments, including technological advances, globalization, a loss of manufacturing employment, and the decline of unions, have reduced the number of jobs that require limited skills but pay a middle-class wage — the very kind of jobs that once moved poorer Americans into the middle class.

Finally, changes in partner selection have also widened the opportunity gap. Not only do those from better-off families tend to end up with more schooling and higher-paying jobs; they are more likely than ever to marry (or cohabit with) others like themselves, according to research by the sociologists Christine Schwartz and Robert Mare.

For all these reasons, the gap in opportunity between the United States’ rich and poor has expanded in recent decades. Left unchecked, the trend threatens not only to offset the progress the United States has made on gender and racial equality but also to usher in a future of deep and hardened class divisions.

It might be tempting to shrug and conclude that the high and increasing opportunity gap in the United States is an unfortunate but inevitable consequence of economic and social shifts. The problem with this reaction is that other affluent democracies do better. The United States has lost its historical distinction as the land of opportunity. Yet there is at least some good news: the fact that other countries are more successful in this area suggests that with the right policies, the United States could do better, too.


One simple, straightforward solution would be to get more money into the hands of low-income families with children. The education policy experts Greg Duncan, Ariel Kalil, and Kathleen Ziol-Guest have found that for children who grew up in the United States in the 1970s and 1980s, an increase in family income of a mere $3,000 during a person’s first five years of life was associated with nearly 20 percent higher earnings later in life. The finding suggests that government cash transfers of just a few thousand dollars could give a significant lifelong boost to the children who need it most. Most other affluent countries, including those that do better on equality of opportunity, offer a universal “child allowance” that does exactly this. In Canada, for instance, a family with two children receives an annual allowance of around $3,000, and low-income families with two children might receive more than $6,000. The United States has only a weaker version of the benefit, the Child Tax Credit, which doles out a maximum of just $1,000 a year per child. Moreover, receipt of the money is contingent on filing a federal tax return, which not all low-income families do.

Other solutions involve Washington getting involved in home life. Fewer children in the United States grow up with both biological parents than in any other affluent country for which data are available. To remedy this, some, such as Barbara Dafoe Whitehead and David Popenoe, co-directors of the National Marriage Project at Rutgers, favor efforts to promote marriage. But research by the sociologists Kathryn Edin, Sara McLanahan, and Paula England and others suggests that this strategy is misplaced. Since women today need less from marriage and expect more from it than they used to, those who are better educated and better off tend to take more time to get established in their jobs and find good partners, which enhances the likelihood of a lasting marriage (or cohabitation). They delay childbearing as well. Among poorer and less-educated women, who see little prospect of a fulfilling and lucrative career, having a child in their teens or early 20s remains common. These women are less likely to stay with a partner: they have had less time to mature personally and to find a person with whom they are compatible, their partners are more likely to have weak financial prospects and a preference for traditional gender roles, and the presence of a child heightens financial and interpersonal tensions. Given all this, convincing more young low-income couples who get pregnant to marry is unlikely to produce many lasting relationships.

Genuine progress probably hinges on poor or less-educated women delaying childbirth. Eventually, this will happen; the teen birthrate has already been dropping for nearly two decades, albeit slowly. For its part, Washington (or any other government) has only limited tools to speed it up. The best might be an education campaign, as Ron Haskins and Isabel Sawhill, policy experts at the Brookings Institution, have suggested, that focuses on the benefits of the “success sequence”: first education, then a stable job, then marriage, and then children.

What about parenting practices, which have a clear effect on childhood development? Although few Americans support extensive government intrusion into home life, one potentially acceptable way that Washington and state governments could try to improve parenting is by paying for home visits by nurses or counselors and providing free or low-cost parenting classes. Getting people to change their behavior and routines is very difficult, so the benefits of such programs are inevitably modest. Nonetheless, in a recent review of existing research, the sociologist Frank Furstenberg found evidence that programs aimed at teaching better practices to parents of children at middle-school age or younger yield some improvements in school readiness and school performance.


Given the difficulties of altering home life, improving schools remains the United States’ main tool for assisting less-advantaged children. For all their inadequacies, public schools do help equalize opportunity by improving students’ cognitive abilities. During summer vacation, the cognitive abilities of children in low-income families tend to regress, relative to those of their more advantaged peers. In other words, these children would lag even further behind if they never attended school. 

A universal system of affordable, educational child care and preschool could help close the capability gap that opens up during the early years of life. Additionally, it would facilitate parents’ employment and thereby boost household incomes, making it doubly helpful for children in low-income families. The Nordic countries offer some lessons: in the 1960s and 1970s, these countries introduced paid maternity leave and publicly funded child care. Today, early education teachers there have training and pay comparable to those of elementary school teachers. The cost of early education is capped at around ten percent of household income. In all these countries, a person’s cognitive abilities, likelihood of completing high school and college, and eventual success in the job market tend to be less heavily determined by his or her family’s wealth and makeup than in the United States.

There has been some movement to expand the United States’ child-care and educational systems at the state level in the past two decades. Most states now have full-day public kindergarten, and some have added public preschool for four-year-olds. But the progress has been very slow, and in recent years, it has been set back by state revenue shortfalls. Assistance from Washington would be of considerable help.

The equalizing effects of college, too, cannot be overstated. Among Americans whose family incomes at birth are in the bottom fifth but who get four-year college degrees, 53 percent end up in the middle fifth or higher. That is pretty close to the 60 percent chance they would have with perfectly equal opportunity. Washington needs to do better at helping people from less-advantaged homes afford college. The average in-state tuition at an American four-year public university exceeds $8,000. In Norway, Sweden, Denmark, and Finland, attending four-year public universities is free. According to data from the Organization for Economic Cooperation and Development, in those nations, the odds that a person whose parents did not complete high school will attend college are between 40 and 60 percent, compared with just 30 percent in the United States.


Employment is the next challenge. First, the low-hanging fruit: since a prison record impedes labor-market success, the United States should rethink its approach to punishment for nonviolent drug offenders. According to the sociologist Bruce Western, states that have reduced imprisonment over the past decade, instead turning to alternative punishments, such as fines and community corrections programs, have experienced drops in crime similar to states that have increased imprisonment. If other states were to follow suit, the United States could avoid needlessly undermining the employment opportunities of a significant number of young men from less-advantaged homes.

Broader trends in the labor market since the 1970s present a stickier problem. Hourly wages at the median and below have not budged in inflation-adjusted terms. In the 1980s and 1990s, the United States created a lot of new jobs. These facilitated the movement of women into the work force and thereby helped many households enjoy rising incomes despite the stagnation in wages. But in the early years of this century, employment growth stopped, and the subsequent recession and slow recovery have dealt a crushing blow to the less skilled. The employment rate among men aged 25-54 who did not finish high school dropped by ten percentage points between 2007 and 2010.

Eventually, the U.S. economy will get back on track, but that will not automatically lead to more jobs and higher wages. The lone period of sustained wage growth at the middle rung and below occurred in the late 1990s. What distinguishes that period is that the Federal Reserve allowed the unemployment rate to drop to four percent, well below what many economists believed to be the level at which inflation would accelerate. If and when the United States returns to low unemployment, it will need the Federal Reserve to again be willing to allow wages to rise significantly before stepping on the brakes.

It would be foolish to count on this, though, so the United States would do well to consider alternative strategies. One useful tool might be the Earned Income Tax Credit. At the moment, the EITC provides an annual subsidy of up to $6,000 to households with less than $50,000 in earnings. That is helpful, but for a person with no children, the credit amounts to less than $500. That group — young adults with low earnings and no children — includes many Americans who grew up in disadvantaged circumstances. If the economy is growing but wages are not, the United States can and should offer a bigger boost to these people’s incomes.

In the past year, a number of commentators, most notably Alan Krueger, chair of the White House Council of Economic Advisers, have suggested that reversing the rise in income inequality could improve economic mobility in the United States. After all, among the countries for which there are comparable data, those with less income inequality tend to have higher relative intergenerational mobility. The United States was already on the high end of the income-inequality scale a generation ago, and since then it has moved even further in that direction.

Yet general calls to reduce income inequality offer little help in identifying which policies to pursue. Consider three possibilities. First, imagine that Washington legislated a radical reduction in the pay differentials for various types of jobs. (Narrower pay differentials account for part of the smaller opportunity gaps in the Nordic countries.) This certainly would reduce income inequality. It would also reduce opportunity inequality: at least in the first generation, even if someone’s capabilities matched perfectly those of his or her parents, his or her income would not. But such a drastic step is not likely to happen, in part because few Americans would support it. Second, suppose the United States were to raise income tax rates for the top one percent of households and lower them for middle-class households. Such a move would reduce income inequality, but it would do little to improve the opportunities of children in low-income families. Third, suppose the United States increased tax rates for all households and used the revenue to fund universal early education. (As the political scientist Andrea Campbell recently wrote in these pages, most other advanced democracies devote far more tax revenue to social programs.) That step would do little to counter income inequality, but it could substantially expand opportunity. A reduction in income inequality, in short, is neither necessary nor sufficient for achieving a reduction in inequality of opportunity.


For all that other countries’ experiences can teach the United States, there are also lessons the United States should take from its own history. The most direct way that Washington has made opportunity more equal in the past has been through affirmative action. Affirmative action is not a strategy that many other affluent countries have embraced, but it has a proven track record in the United States. Since the late 1960s, affirmative action programs for college admissions and for hiring have expanded opportunities for women and various minority groups.

Now, a number of observers from across the partisan spectrum, from Richard Kahlenberg, a senior fellow at the left-leaning Century Foundation, to Charles Murray, a fellow at the right-leaning American Enterprise Institute, favor shifting the focus of affirmative action efforts from race and gender to family background. Emphasizing family background would continue to disproportionately help African American and Latino children, since they are more likely to come from families with low incomes and other disadvantages. Indeed, it would do more to help poor black and Latino children than traditional race-based affirmative action programs, which have mainly benefited middle-class members of such minority groups.

In response to court rulings and ballot initiatives outlawing consideration of race in admissions decisions, some public university systems, including those of California and Texas, have already moved in this direction. One approach guarantees the top ten percent of students graduating from any public high school in a state automatic admission to a public university in that state. Sometimes, this is helpful; in schools where almost all the students are from poor families, the top ten percent of the graduating class will inevitably include low-income students. A more direct strategy would be for colleges and universities to consider family background as one of several kinds of disadvantages that applicants may have faced and to include that among the criteria by which applicants are ranked.

How might employers be persuaded to use this direct approach? Half a century ago, the federal government mandated the use of affirmative action in public agencies and in firms with which it contracted. It could do the same now in order to address the nation’s new opportunity gap.

In the last half century, the United States has taken long strides toward equalizing economic opportunity. That progress did not happen on its own; it took place with a push from the government. In recent decades, however, the opportunity gap for Americans from different family backgrounds has started to grow. Fortunately, the United States’ experience and that of other affluent nations suggest that the country is not helpless in the face of economic and social changes. There is no silver bullet; a genuine solution is likely to include an array of shifts in policy and society. Even so, a fix is not beyond the United States’ reach.

10 thoughts on “America’s opportunity gap

  1. This is a great essay. I agree with 95% of it.

    However, I don’t think I agree with your statement that “universal early education… would do little to counter income inequality.”

    Maybe you mean that universal early education would not immediately and directly reduce income inequality. However, under most plausible estimates, universal early education would have much larger percentage effects on future incomes for households who would be expected to have lower future incomes.

    I elaborate on this point and provide more evidence in a post at my blog:

  2. So what we should really do is remove children from lower socio-economic backgrounds where parenting skills, for reasons discussed by the author, are inferior and place them in superior, higher socio-economic environments? Lots of upper middle to upper quintile familes would like more children so, per the authors logic, this would improve their outcomes.

  3. TIm: You’re right that I was referring to the near-term. But I’ll say I’m increasingly skeptical about the long-run too. Forces working against wage growth for those in low-end jobs are now very strong — intense competition (both foreign and domestic), technological change, demands on management for strong quarterly profits, weak unions, growing use of pay-for-performance, and others. We may now live in a world in which pushing up the low end of cognitive skills and noncognitive traits is a good thing on opportunity-mobility grounds but, unlike in the period from 1870 to 1970, doesn’t raise the earnings of whoever ends up in the low-end jobs. There’s the additional matter that since the mid-1990s the rise in income inequality in the U.S. has been mainly a 1%-99% story, which universal early education probably won’t affect.

  4. Lane: I now understand your point better. I agree that early education won’t do much about the gap between the top 1% and the bottom 99%. However, I am more optimistic that the overall wage structure and jobs structure would respond to some degree to a significant skill upgrading of a broad range of the population.

  5. “The lone period of sustained wage growth at the middle rung and below occurred in the late 1990s. What distinguishes that period is that the Federal Reserve allowed the unemployment rate to drop to four percent, well below what many economists believed to be the level at which inflation would accelerate.”

    This is a small point in your piece, but can you explain to readers how the Federal Reserve can influence the unemployment rate?

  6. It is an interesting conundrum. There are too many variables and unknowns I worry and fear all the do-gooders. It is clear we created the problem through our ignorance

  7. The late 90’s were driven by a tech boom led by the DotComs. That proved to be almost as large a fraud as the housing boom of the mid-2000s. Billions made and spent followed by billions lost.
    I find the other points about ever more high impact parenting among higher income families very interesting. There are two ways to grow a gap and one is for the leaders to go ever faster rather than the laggards slowing down much. Hard to stop that in a free society.

  8. “Fewer children in the United States grow up with both biological parents than in any other affluent country for which data are available.”

    If you have a valid citation for this statement, please provide it.

    The US is a leader but almost certainly not THE leader in this category. See OECD surveys for example. Also:

    “Countries with a higher proportion of births to unmarried mothers include Iceland, Sweden, Norway, France, Denmark and the United Kingdom.”

  9. Re: “Genuine progress probably hinges on poor or less-educated women delaying childbirth. Eventually, this will happen; the teen birthrate has already been dropping for nearly two decades, albeit slowly. For its part, Washington (or any other government) has only limited tools to speed it up. The best might be an education campaign, as Ron Haskins and Isabel Sawhill, policy experts at the Brookings Institution, have suggested, that focuses on the benefits of the “success sequence”: first education, then a stable job, then marriage, and then children.”

    I’m extremely skeptical than a public education campaign of this sort would have any real-world impact on the target audience (people in their late teens and early 20s). Do we have any evidence at all suggesting that it would? Have, for example, the “Married People Earn More” billboards in Baltimore and other cities had any positive effect on marriage? I’m guessing not.

    Concrete policy proposals that focus on the material, institutional, and programmatic contexts of disadvantaged young people (ages 18-24) seem like a much better bet. One example, working adults under age 25 are only eligible for the EITC if they have one or more children. This seems like very bad policy if we want to encourage stable employment among young people before they have children.

    Proponents of a “success sequence” public education campaign might counter that even if it doesn’t work, it wouldn’t have any negative effect (other than the waste of public funds that could have been better used for things like extending the EITC to young people in low-wage jobs), so it’s worth trying. But if it hardens “just world” biases in a way that make it harder to implement progressive policies that increase opportunity and security for people who don’t follow that sequence, it actually could be detrimental. Note here that Rick Santorum was the only Presidential candidate to run with the data used by Sawhill/Haskins to support their proposal.

    Finally, I’ve written at the CEPR blog about the data used by Sawhill/Haskins (and highlighted by Santorum and Ross Douthat). See: Basically, I think they really overstate the strength of their evidence (particularly when it comes to causation).

Leave a Reply