Vol 13, No.2 (2003)
ISSN 1546-2250
Doing Poorly: U.S. Child Poverty in Cross-National Context1
Lee Rainwater
Luxembourg Income Study
Harvard University
Timothy M. Smeeding
Maxwell's Center for Policy Research
Syracuse University
Citation: Rainwater, Lee and Timothy M. Smeeding. “Doing Poorly: U.S. Child Poverty in Cross-National Context.” Children, Youth and Environments 13(2), 2003. Retrieved [date] from http://colorado.edu/journals/cye
| Comment on This Article |
Abstract
This article compares the levels and patterns of child poverty in the United States to those found in 14 other rich nations in Europe, Canada and Australia. It examines demographic variations in child poverty rates, and analyzes the roles of both family earnings and social transfers as either contributors to or reducers of levels of child poverty. Since the poverty rates of children in two-parent versus single-parent families differ greatly, separate sections of the paper describe the “income packages” and the patterns of child poverty for each family type. Finally, we consider policy options for reducing the very high rate of U.S. child poverty.
Keywords: poverty, inequality, children, social policy, single parents
For some 40 years now poverty has been a central and self-conscious concern in U.S. society. The War on Poverty, officially launched in 1964 by President Lyndon B. Johnson, spawned a large research establishment and literature (Johnson 1964). As analysts have dug into this large issue it has proved fruitful to investigate the special circumstances and dynamics of different groups of the poor, particularly the notably dependent populations of the elderly and the children. This paper is about the economic well-being of the latter group. It examines U.S. children from a particular perspective, that of their fates in comparison with the fates of children in some other rich nations.
Comparing the situation of U.S. children in low-income families with their counterparts in some other rich countries gives us a powerful perspective on the dynamics of child poverty in the United States in the 1990s. The analysis provides not only detailed examinations of the patterns of child poverty in the United States but comparisons with child poverty in 14 other countries, 12 of which are in Western Europe—Belgium, Denmark, Finland, France, Germany, Italy, the Netherlands, Norway, Spain, Sweden, Switzerland, and the United Kingdom—and two of which are the Commonwealth nations of Australia and Canada. The surveys on which our analysis is based derive from the Luxembourg Income Study, a database available to social scientists throughout the world that contains nearly one hundred income surveys.2
Thinking, Rethinking Poverty
For much of the time since it was launched, poverty warriors have proceeded as if the War on Poverty were a new enterprise in U.S. society, a project designed to eradicate the vestiges of poverty left after rapid economic growth in the post-World War II period. More recently, several excellent historical studies of the way U.S. society has dealt with economic marginality and disadvantages have emphasized the continuity of issues and strategies in the War on Poverty with previous social welfare politics and policy (Ruggles 1990; Katz 1986; 1989; Patterson 1981).
Postwar America's concern about poverty has long roots. The recognition that economic growth does not eradicate poverty, dating from the eighteenth century, may be relatively modern, but concern over the impact on society of large classes of poor people predates economic strategies for dealing with it. Indeed, poverty has been a subject of concern, debate, moralizing, and controversy throughout written history. Many would say today as in biblical times: “The poor shall never cease out of this land.”
Economic and political strategies counter such fatalism by raising quantifiable questions, such as: How large is the class of the poor? Even though it is a rare society that would assert that it has no poor people, there may still be important differences among societies in how many poor they include. Surely it is important to count the number of poor people in a society, since we judge societies very differently when we discover that in some there are a great many poor people and in others only a few.
In highly stratified societies, the economic resources available to individuals and families vary greatly. As members of these societies, we develop a lively sense of the implications of different levels of resources—that is, we see that experiences of plenty or of deprivation often seem to be critical in determining which other members of society we choose to interact with and the kinds of social participation that are feasible for us. In short, if we do not look at poverty as one aspect of socioeconomic inequality in modern societies, we are missing its essential nature. Any correct study of poverty must proceed from a sociologically grounded understanding of the interpenetration of material and social well-being in modern societies, recognizing also the particular degree of socioeconomic inequality in the social stratification system of each society.
The definition of poverty adopted by the European Council of Ministers in 1984 reflects such a concept of poverty. The European Union (EU) member countries now ground their approaches to poverty in an understanding of the nature of social stratification in prosperous industrial and postindustrial societies such as their own: “The poor shall be taken to mean persons, families and groups of persons whose resources (material, cultural and societal) are so limited as to exclude them from the minimum acceptable way of life in the member state in which they live” (European Commission 1985).
By contrast, writings about poverty in the United States tend not to focus so sharply on social exclusion but rather on the characteristics of the poor themselves. As Michael Katz (1989) argues, Americans have always been deeply preoccupied with the distinction between the undeserving and the deserving poor. (This concern has also been characteristic of Britain and its Anglo-Saxon Commonwealth nations.) In the United States the struggle to define the poor in these terms has been centered most recently in the welfare reform controversy, which has resulted in the abolition of the program Aid to Families with Dependent Children (AFDC). The undeserving/deserving dichotomy has now produced a welfare program, Temporary Assistance for Needy Families (TANF), the premise of which the late Senator Daniel P. Moynihan (1996) characterized at its passage as that “the behavior of certain adults can be changed by making the lives of their children as wretched as possible.”
Within this orientation, two perspectives contend in U.S. characterizations of poverty. One sees economic status—that is, people's command over goods and services (Ruggles 1990; Citro and Michael 1995)—as the defining and sufficient indicator of being poor. The other perspective is more broadly concerned with people's socioeconomic situation. It was the latter kind of sociological concern that dominated the elite interest in poverty in the 1960s and launched the War on Poverty. Michael Harrington's The Other America: Poverty in the United States, first published in 1962, is generally credited with putting poverty on the agenda of the administration of President John F. Kennedy. Harrington combined his own experience as a writer about the U.S. working class with much sociological research during the 1950s on the inner city, juvenile delinquency, and slums. For Harrington, there was
a language of the poor, a psychology of the poor, a worldview of the poor. To be impoverished is to be an internal alien, to grow up in a culture that is radically different from the one that dominates society....[The poor] need an American Dickens to record the smell and texture and quality of their lives. The cycles and trends, the massive forces, must be seen as affecting persons who talk and think differently (Harrington 1962, 18).
Social Measurement of Poverty
Although the U.S. official poverty line is considered an absolute one, in fact no one has ever argued that it is unchangeable. It is understood that there is a historic component to the definition of a minimum standard of living; for instance, the U.S. poverty line would obviously be a line of affluence in many other parts of the world. Even so, some argue that it is possible to establish, for the time being, an absolute standard, and that it is unnecessary to adjust the poverty line regularly for increases (or decreases) in average incomes. At the core of the definition of an absolute poverty line such as the Orshansky index, however, is a hazily recognized contradiction—poverty is actually relative, but for now we will define it as unchanging in real terms (see also Smeeding, Rainwater, and Burtless 2001).
The more experience countries have with absolute poverty definitions, the more obvious becomes the absurdity of the rationale for them. Not surprisingly, attempts have been made to revise poverty lines so as to avoid this underlying absurdity. Any such effort requires that poverty lines be adjusted to reflect changes in living standards—an approach that has been exhaustively pursued, for example, by a committee of the U.S. National Research Council (Citro and Michael 1995).
A social measure of poverty is concerned ultimately not with consumption but with social activities and participation. Researchers with this orientation do not look at the problem of poverty in relatively affluent societies as one of low consumption per se but focus instead on the social and personal consequences of poor individuals' inability to consume at more than an extremely modest level. Without a requisite level of goods and services, individuals cannot act and participate as full members of their society, and it is this participation in social activities that confers utility, not consumption. While such a view is mainly identified with sociological (and anthropological) traditions, a few economists have adopted its insights by focusing on consumption as an intermediate activity—that is, as an input to social activities that in turn confer utility (Duesenberry 1949; Lancaster 1971; Sen 1992).
Denton Vaughan provides an excellent description of the social meaning of poverty in contemporary societies:
[I]n the complex, largely urban, and industrial and service societies of the post World War II United States, Canada and Western Europe, the poverty problem stems from the existence of substantial population subgroups whose members lack the material resources required to perform—except with the greatest difficulty—roles in the central societal domains of family, work, and citizenship as defined by the mainstream members of society and as generally accepted by members of the low income groups themselves. [Members] have a relatively well-developed sense of the material resources associated with different levels in the material status hierarchy. It is this sense that permits the individual to judge the difference between a good salary and a poor one, a nice car and a bare bones econo-box, or a decent apartment and a slum tenement, and more generally to assess his or her location in the overall stratification system (Vaughan 1993, 22-23).
As Vaughan makes clear, we identify someone as poor by not only looking at social conditions but taking into account related social identities and meanings—those associated with both the poor person and, even more, the other members of society who hold that person responsible for his own social position.
We present first an overview of U.S. child poverty compared with child poverty in 14 other rich countries. To sample a broader range of the comparison possibilities, we also discusses poverty rates among 25 other rich and not-so-rich nations and then set the child poverty rates of the 15 rich countries in the context of income inequality more broadly. Finally, we examine the role of income packaging and the role of different forms of income in alleviating child poverty for all families and in more detail for families headed by a mother alone. (A more detailed analysis of the issues dealt with here and the methodological choices that have to be made in conducting such a study can be found in Rainwater and Smeeding 2003.)
It should be evident from the discussion so far that although poverty analyses make use of a poverty line there is no hard and fast qualitative difference between being poor and not being poor. The ability to participate meaningfully in one's society is not an all-or-none thing. The contemporary social standards for what constitutes poverty or near poverty or mainstream living are nodes along a continuum of affluence. When we use income as an index of material well-being we can conveniently rank people along a continuum from very low income to very high income. The poverty line, and our counts of those below it, is a snapshot of part of that continuum in order to facilitate comparison—among groups, over time, among nations. Behind this simplification lies the complex reality of people's daily activity of converting their material resources into some sort of social participation. We want to count the poor as those for whom there must be a great struggle to thus convert their meager resources.
Child poverty is not a rare occurrence in the United States or in the other countries to which we compare it, although we will see there are large differences among countries in the proportion of children who are poor. By a convenient coincidence, around the end of the century there were roughly the same number of children in the United States and the 12 European countries—around seventy-two million. But we find as many as fourteen million poor U.S. children compared to some eight million poor children in our comparison European countries. Both are large numbers, but the United States certainly has a larger child poverty problem than this group of European countries.
In defining a poverty line one must establish indicators of the various aspects of economic well-being. How do we index (a) the economic resources of families, (b) the needs of the family as a function of who its members are, and (c) the standard of living against which to compare a family's resources?
The data available in the Luxembourg Income Study database allow us to index economic resources by the after-tax family income. In that income we include the money the family receives from earnings, assets, and transfers, and some “near cash” income in such forms as food stamps, housing allowances, or tax credits like the Earned Income Tax Credit. Because of difficulty in measuring their incidence and their value we cannot include important non-cash resources available to some families—for example private or public health insurance, or mortgage subsidies.
We divide after-tax income by a need factor based on the number of persons in the family and the age of the head. The result is equivalent income, which we take as an index of dissimilar families' ability to maintain a given level of economic well-being. Obviously, the more people in the family the more resources it needs to maintain a given level of well-being. But because there are economies of scale, need does not increase proportionately with size. We have also found evidence that social need increases with the age of the family head to a point in middle age and then declines, so we also adjust for age of head.
For an index to the mainstream standard of living in the 15 countries we are analyzing, we calculate the median equivalent after-tax income of the individuals in the LIS sample, and take one-half of that median for the poverty line. We count the economic well-being of families as the ratio of their incomes to the median. We count families and individuals as poor when equivalent income is less than half of the median.
It will also be helpful in the analysis below to look at the income groupings of children's families: high (one-and-a-half times the median), middle (around the median), marginal or low (two-thirds of the median), or below the poverty line (half of the median). Some examples will give an indication of what these income classes represent in the United States. We label as high-income persons living in a family of four (with a head 45 years of age) whose 1997 after-tax family income was $66,700 or higher. A family with an income $44,455 a year is exactly in the middle, and individuals are defined as low-income if their family income was below $29,600 a year. The U.S. LIS poverty line in 1997 is drawn at $22,227 a year.
To give an idea of where poverty starts, Table 1 shows the approximate U.S. LIS poverty lines in the 1997 data for families of different sizes (all with a head 45 years of age).
Table 1. LIS Poverty Lines for the United States, 1997
Table 2 shows the median equivalent incomes in each country after the family size and age adjustments, as described above. In the analysis that follows, persons are categorized in different income classes depending on where they stand in relation to their national median equivalent income.
Table 2. Median Equivalent Income in 15 Countries in the 1990s
Poverty Rates of All Children
Looking first at child poverty rates in the 15 countries (see Figure 1), we find that the United States has an extremely high poverty rate of 20 percent. That is, 20 percent of all children live in families with equivalent incomes below one-half of the median equivalent income. These are incomes so low that the children and others in their family are not able to participate enough in community activities to be perceived, by both themselves and others, as regular members of society.
Italy has the second highest child poverty rate at just below 20 percent, followed closely by the United Kingdom at 16 percent. The two Commonwealth nations, Canada and Australia, have rather high rates at more than 13 percent, and at 12 percent Spain's child poverty rate is almost as high.
The poverty rates in the rest of the countries, starting with Germany, are all below 10 percent. France has a rate of 7 percent, as does the Netherlands. Switzerland follows at 6 percent. The rest of the countries have very low rates of 5 percent or less. Very low rates of child poverty prevail in the four Nordic countries.
In short, the poverty rate of U.S. children is over four times as high as that of children in northern Europe. The rate in the United Kingdom has a disparity that is almost as dramatic—British children suffer a poverty rate over three times as high as that of the northern European countries.
What are the odds of escaping poverty in these different countries? Answering that question, we see even starker contrasts. A U.S. child has a slightly less than four-to-one chance of escaping poverty, and the odds are about the same for an Italian or British child. In contrast, a Nordic child has a twenty-five-to-one chance or better of escaping poverty. These children's chances of not being poor are six or more times greater than those of a U.S. child, and around five times greater than those of a British child.3
If we broaden our focus to all of the countries for which LIS data exist, we find no wider range in poverty rates, but some interesting possibilities for comparison do emerge when we consider the poverty rate for the most recent years for each country (Table 3). Two Western European countries—Austria and Luxembourg—fit neatly into the group of Northern continental countries with rates under 5 percent. Ireland's high poverty rate of 12 percent, on the other hand, is about the same as Spain's, Australia's, and Canada's, but not as high as the 16.3 percent rate in the United Kingdom.
Table 3. Trends in Child Poverty 1970-1997
Many people would choose not to include data from the five Eastern European countries under the rubric of rich nations, but we should keep in mind that their standards of living are not much lower than those of Western Europe some 40 years ago. By world standards, they certainly are rich nations.
The most interesting observation to be made about these five Eastern European countries is that the range in child poverty in these countries is as great as it is in the countries of Western Europe. In fact, of all the 25 countries, the two with the very lowest child poverty rates are the Czech Republic and Slovakia. Their rates are even a bit lower than those of Sweden and Finland. Hungary and Poland appear in the middle of the range with a somewhat higher rate than France's but lower than Spain's. At 23.2 percent, the child poverty rate in Russia is higher even than that of the United States.
Data are available for two other countries—Israel and the Republic of China (Taiwan). With rates not too different from the German poverty rate but lower than those of the two Commonwealth nations, the United Kingdom, and the United States, these countries have generally higher rates than most European countries.
Thus broadening our scope from 14 to 25 countries tells us that, within the range of economic development that these countries represent, there is little relationship between a country's prosperity and its child poverty rate. The West European countries with very low poverty rates have gross domestic products close to three times those of the Czech Republic and Slovakia, and yet their poverty rates are not very different. Spain and Ireland have twice the real GDP of Poland, but the three countries have very similar poverty rates. The United Kingdom, Italy, the United States, and Russia represent an enormous range in GDP, yet these four all have very high poverty rates.
If we focus on just the 15 countries of primary interest, we find that the correlation between GDP per capita and poverty rates is effectively zero. As suggested by the broader scope, there is simply no association between how rich a country is and how unequal its equivalent income distribution is. National riches and national child poverty are neither positively nor negatively associated.
Trends in Child Poverty
The results examined so far give us a snapshot of child poverty rates in the 1990s that depicts most prominently the very large differences among countries. Has it always been thus? Have these differences been stable over time, or are some countries experiencing increasing or decreasing rates?
To answer these questions, it would be desirable to have solid evidence gleaned from regular surveys over several decades for many countries. Unfortunately, there is no such database, and it is unlikely that such a database could even exist for periods before the 1970s. However, the LIS database does have data from 1970 on for a few countries, and for more countries since the 1980s. Thus, we have observations for some countries over a 25-year period and for others over periods up to 15 years. Because we have sometimes had to use different surveys at different points in time, and because sometimes the same survey has changed its methodology from one time to another, we can feel fully confident of a change only when there is a clear trend across more than two surveys, and preferably across more than three. When there is a jump solely from one period to the next, we have weak evidence of a change.
Reexamining Table 3, we find no country in which the child poverty rate has declined significantly. In most countries there seems to be considerable stability in the child poverty rate. Two countries, however, have experienced large increases in child poverty. In the United States we find a steady increase from 1969 (13.1 percent) through 1986 (22.9 percent) and then a leveling off—to 22.7 percent in 1994 and 20.3 percent in 1997. Child poverty, in short, increased by more than 50 percent in the United States over the last quarter-century.
It is interesting to note that even though the rate of U.S. child poverty around 1970 was a little lower than Canada's rate, it rapidly surpassed Canada's in the 1970s, and the difference between the two countries continued to grow until the mid-1980s. By the mid-1990s the odds of a Canadian child avoiding poverty were almost twice as great as the odds for a child in the United States.
We must temper our conclusions for the United Kingdom. Although the survey is the same in all years, the rules and procedures for calculating disposable income may well have changed over the 25 years since 1969, and therefore we cannot be sure that the six surveys are exactly comparable. Even so, it is highly unlikely that the changes we observe are not pretty close to what actually happened. Our results are consistent with those of other British studies of income inequality.
It would appear that since 1969 there has been a steady increase in child poverty in the United Kingdom, from around 5 percent to the most recent observation of 16 percent. In the United Kingdom, then, child poverty seems to have tripled. In the early 1970s the child poverty rate in the United States was two and a half times that of the United Kingdom, but by the 1990s it was only one-quarter higher.
We can assume for almost all of the countries for which we have sufficient data that the levels of child poverty are quite stable characteristics of their societies. Only for the United States and the United Kingdom do we have clear evidence of major change in the economic well-being of children. The United Kingdom has moved from being a country with low child poverty to one with a very high rate. Likewise, the child poverty rate in the United States has gone from high to very high. In contrast, four countries with moderately high child poverty rates—Spain, Israel, Australia, and Canada—have shown no significant increases or decreases.
Thus, the picture of child poverty in rich countries in the latter part of the twentieth century is largely static. While we find marked increases in three countries, there are no countries in which the proportion of children experiencing the social exclusion attendant on low incomes has significantly decreased. The effects of income-producing institutions on children seem not to have changed in an important way anywhere except in the United States, the United Kingdom, and Italy.
The previous section outlined the broad picture of child poverty in the 15 countries of interest. Now we can turn our attention to the need to understand poverty as one aspect of the overall inequality in economic well-being in a society. From the perspective of income distribution, we can think specifically in terms of income classes ranging from extreme poverty to great riches. The three income groupings described above—high, middle, and marginal or low—can be subdivided more finely into seven income classes:
High-income: More than one and a half times the median equivalent income
• Very rich: More than three times the median
• Rich: More than twice the median
• Prosperous: More than one and a half times the median
Middle-income: Between two-thirds and one and a half times the median
Marginal income: Less than two-thirds of the median
• Near poverty: Less than two-thirds of the median
• Poverty: Less than half of the median
• Extreme poverty: Less than one-third of the median
While the income classes into which we have divided the income distribution are arbitrary they allow one to grasp the significance of differences in distribution more intuitively than would an overall measure of distribution. Figure 2 shows the percentage of children in the low-, middle-, and high-income classes in each country. The dark bar, representing the size of the middle-income group, is very large—more than three-quarters—in the Nordic countries and the Netherlands. The middle-income children in Germany, Switzerland, and Belgium number at least 70 percent. France follows, with two-thirds of its children in the middle-income group. Spain, Canada, and Australia are clustered, with around 60 percent of their children in the middle-income group. Fewer than half of children are middle-income in the United Kingdom, Italy, and the United States.
Figure 2. Marginal, Middle and High Income Children
There is a positive correlation between the proportion of children in the high- and the marginal-income groups (.81). In the United States there are almost as many prosperous and rich children (17 percent) as there are poor children. The same is true in the United Kingdom and Italy. The countries with very low poverty rates also have low proportions of children living in prosperous and rich families—around 10 percent or less.
Table 4 shows the distribution of children in the seven income classes. Focusing more sharply on the marginal-income group, we observe that it ranges in size from 7 percent in Sweden to 34 percent in the United States, the United Kingdom, and Italy. The distribution of extremely poor children in the United States is even more distinctive. Over 9 percent of U.S. children and nearly 11 percent of Italian children are extremely poor, while in all other countries the extreme poverty rate is 5 percent or less, and in seven it is less than 3 percent. The contrast with the United Kingdom is also striking—fewer than one-third of English poor children are extremely poor, while about half are extremely poor in Italy and the United States.
Table 4. Percent of Real U.S. GDP per Median Equivalent Person
Considering the marginal-income group as a whole, we note that in most countries the great majority of those with marginal incomes are in the near poor rather than the poor group. The exceptions are Italy and the United States, where only a minority are near poor. In the Nordic countries, Belgium, and Switzerland, two-thirds or more of the marginal-income children are near poor rather than poor. In the remaining countries, around half of the marginal-income children are near poor.
As will be apparent below, the social protection transfers provided by the government for lower-income children in the United States are inadequate to protect them from poverty, whereas such protections are dramatically greater in Europe.
Up to this point we have been discussing poverty and income distribution based on a single measure of income—total money (and near cash) income minus taxes, adjusted for family size. Because we subtract taxes we are also taking into account the effect of tax rates and tax expenditures on disposable income. Now we move on to a discussion of income packaging and child poverty.
Income packaging refers to the fact that a family's income is most often in most countries an aggregate of several different sources. An extremely detailed accounting of families' incomes in a particular country might require attention to as many as 50 different sources of income—different members' taxes, earnings, rent, interest, dividends, child support, and a vast array of social benefits such as pensions, child allowances, public assistance, housing allowances, and many other particular transfer programs. Each source of income involves a claim and entitlement based on some activity or characteristic of family members—employment, capital, a social insurance contribution history, low income, etc. In the analysis of income packaging we look at three main aggregates of income—market income, taxes, and transfers. (For background on the concept of income packaging see Peattie and Rein 1983; Rainwater, Rein, and Schwartz 1986.)
What are the roles of various sources of income—particularly earnings and the range of social protection transfers provided by governments—in enabling families with children to avoid poverty?
Income Packaging and the Escape from Poverty
Pre-transfer income includes salaries, self-employment income, and asset income before taxes are paid. Post-transfer income (disposable income) refers to the amount left after certain interventions by government, which collects income taxes and social insurance contributions and makes social protection transfers. Post-transfer income, in other words, is the net of pre-transfer income minus taxes plus social and private transfers. Government redistributes income by taxing pre-government income and paying part of the taxes it collects to citizens in the form of social transfers. (Think of taxes and social insurance contributions as a negative transfer.) It also mandates various private transfers by law or through the courts.4
We compare in Figure 3 the poverty rates of children based on their families' pre-transfer income with the rates we have examined above based on post-transfer income. The figure graphs three elements of the redistributive process. The horizontal axis shows what the child poverty rate would be if families had only their pre-transfer incomes. The vertical axis shows the poverty rates we are already familiar with, which are based on post-transfer income. The rays emanating from zero show the percentage reduction in the pre-transfer poverty rate accomplished by the net of families' taxes paid and transfers received.
Figure 3. Pre- and Post-Transfer Poverty Rates
We note that government reduces child poverty by only a small amount (around 25 percent) in the United States and Italy. The Italian result is somewhat misleading since the child allowances paid by employers as part of earnings are included in pre-transfer income. However, child allowances are not particularly generous in Italy. The reduction is a bit greater in Switzerland, Germany, Spain, Australia, and Canada—between 37 and 46 percent. In sharp contrast, in six countries—Sweden, Finland, Belgium, Denmark, Norway, and France—the net effect of government intervention is to reduce child poverty by 75 percent or more. The United Kingdom and the Netherlands stand in the middle with a little over 50 percent reduction.
A careful look at the seven countries with pre-transfer poverty rates just below and above the 25 percent range shows that, after government intervention, child poverty rates range from a high of 20 percent in the United States and Italy to 12 to 14 percent for Australia, Canada, and Spain, to 10 percent for Germany, to 7 percent in France, and 5 percent and less in Belgium and Sweden. Note that the United Kingdom's pre-transfer poverty rate is much higher than that of the United States, but because its government programs reduce child poverty by over 50 percent compared to 20 percent in the United States, its after-tax-and-transfers poverty rate is lower. Germany has a low pre-transfer child poverty rate, but it nevertheless reduces it further by almost 40 percent.
The calculation of pre- and post-transfer poverty rates is a heuristic device. After all, the redistribution process for the most part takes place simultaneously with the generation of the “primary” market income. Most Europeans in particular are not even sure what their pre-tax salaries are since income tax and contributions are withheld at the source and the yearend tax return is much simpler than the U.S. system. Many European surveys do not even attempt to ask about pre-tax income.
This heuristic is most useful in thinking about the effect of taxes and transfers on the situation of families with earners. For families without earners, there is very little if any pre-transfer income— asset income may be a very small source. But since (with the exception of the United Kingdom) 90 percent or more of children live in families with earners, the pattern of the government effect on poverty is not too different for children in earner families from that in Figure 3 for all children.
The Impact of Family Type on Poverty and Income Packaging
Since the situations of children in single-mother families and in two-parent families differ a great deal in all countries, we consider income packaging as it affects poverty for these two family types separately. It should be said at the beginning, however, that there is little correlation (only –0.10) between the proportion of children in a country in two-parent as opposed to single-parent families and the overall child poverty rate.5
In Figure 4, we see that the proportion of poor children in two-parent versus single-parent families varies from 91 percent of poor children in Italy to 57 percent in single-mother families in Norway. Very few children in any country live in single-father families—the highest proportion is 7.4 percent in Sweden.
Figure 4. Family Types of Poor Children
We note high and low child poverty rates at all points in the range of parent types. The countries with child poverty rates above 10 percent vary from the highest proportions of children in two-parent families (Italy and Spain) to those among the lowest (the United Kingdom and the United States), with Australia and Canada in between. Similarly, there is great variance among those with child poverty rates below 10 percent, from those with relatively high proportions in two-parent families, like Switzerland and the Netherlands, to the countries with middling to high rates of single-parent families, such as the Nordic countries, particularly Sweden.
Levels of Economic Well-being in Two-Parent Families
Table 5 shows the distribution of children in two-parent families by well-being categories. The differences among countries are not as substantial as for all children but are nevertheless quite large.
Table 5. Economic Well-Being Levels of Children in Two-Parent Families
Focusing first on the middle-income group we see a range from 87 percent in Sweden to 50 percent in Italy. We can identify four clusters of countries. The most middle-class countries, all with over 80 percent in the middle-income group, are the Netherlands and the Nordic countries—Denmark, Finland, Norway, and Sweden. They are followed by three countries with middle-class incomes in the 70 percent range—Belgium, Germany, and Switzerland—then three other countries in the 60 percent range—France, Canada, and Australia. Finally we have the four countries—Spain, the United Kingdom, the United States, and Italy—at the low end of the scale with the most unequal distribution of two-parent children's income.
For all of the northern continental European countries, we find close to three-quarters or more of children living with two parents to be in the middle-income group. These are countries with not only relatively low proportions of children who live in poverty or near poverty but relatively few children whose families qualify as prosperous or rich.
We see that very few children living in two-parent families in any country are extremely poor. Italy has the highest rate, with 10 percent. The United States has the next highest level, though only 4 percent of U.S. children in two-parent families have incomes less than one-third of the median. In five countries—the Nordic countries and France—the proportion of very poor children in two-parent families is 1 percent or less. Combining both levels of poverty, we find that only 1 to 3 percent of children in this type of household in the Nordic countries are poor; they are joined by the Netherlands, Switzerland, Belgium, France, and Germany at the very low level of less than 6 percent. In the 8 to 11 percent range we find two-parent children in Canada, Australia, the United Kingdom, the United States, and Spain.
There are interesting variations in the proportion of children in the near poor category (between half and two-thirds of median equivalent income). Italy has a very high proportion of children in two-parent families who are near poor—fully 14 percent—followed closely by Spain and the United Kingdom at 13 percent and the United States at 12 percent. We find a group of European counties—Switzerland, France, Germany, and Belgium—as well as Australia and Canada in the 8 to 11 percent range, while in the Nordic countries and the Netherlands fewer than 6 percent of children in two-parent families are in the near poor group. We turn to the situation of children in single-mother families who are exposed to significantly greater risks that those in two parent families.
Child Poverty in Single-Mother Families
In recent years all countries have given special policy attention to the economic vulnerability of children in single-mother families. Most single mothers are caught between their responsibilities as mothers and heads of families, on the one hand, and the need to earn money on the other. In quite a number of countries there is additional concern that the proportion of children who live in single-mother families may be steadily rising. Such concern has been widespread in the United States, where the increase in single-mother families is perhaps most prominent, but it has also emerged in the United Kingdom and Sweden and among various international agencies such as the OECD.
Our 15 countries show a very wide variation in the proportion of children who live in single-mother families. Only 5 percent of children in Spain and Italy live in such families, while in the United States and the United Kingdom the figure is around 21 percent and 19 percent in Sweden. At 9 percent, Belgium, France, and Switzerland have relatively low proportions of children in single-mother families, and the rate in the Netherlands is lower still, at 7 percent. The number of these children is slightly higher in Germany, Finland, and Australia, at 11 percent, and single-mother families in Denmark, Norway, and Canada number in the 14 and 15 percent range.
The composition of single-mother families can vary significantly across countries, and so it is important to define the term from the outset. Our definition of single-mother family is a family with minor children in which a woman is the head. When other adults are present, as they sometimes are, they are likely to make contributions to the family's income. We find that not only do Spain and Italy have the lowest proportion of single-mother families, but that in over half of such families another adult is present. In the United States there is an even larger proportion of children in families that have both a single mother and another adult present; some 7 percent of all U.S. children live in such families, or almost one-third of children in single-mother families. In the other countries a very small minority of children in single-mother families also have another adult present. Thus, with the exception of Spain, Italy, and the United States, over three-quarters of children in single-mother families in our study live in families in which the only adult present is their mother.
Given the tenuous economic situation of single-mother families, we would expect to find that children in such families are disproportionately concentrated in low-income groups. This is indeed the case. With the exception of Sweden (52 percent), Spain (56 percent), Belgium and Finland (64 percent), and Italy (65 percent), some 70 percent or more of children in single-mother families have incomes that place them in the lower third of the distribution of all persons' income. Indeed, 80 percent or more of such children in the United Kingdom, Australia, Germany, and the Netherlands are in the lower third of the total distribution.
Being in the lower third of the income distribution of all persons does not necessarily mean that one is poor or even near poor. Figure 5 charts the proportion of single mothers' children in the lower third by their poverty rates. Comparing the countries with the highest levels of children in the lower income third, we see that in the United States almost 70 percent of lower-third children in single-mother families are poor, followed by Canada, Australia, and Germany in the mid 50 percent range. In the case of the United Kingdom this is true of fewer than half of the lower-third children living in this type of family. In the Netherlands almost three-quarters of the lower-third children in single-mother families are not poor.
Figure 5. Poverty Rates of Low-Income Single-Mother Children by Percent in the Lower Third
In the Nordic countries, Belgium, and Switzerland, between 12 and 20 percent of the lower-third children in single-mother families are poor, but there is a fairly wide range in the overall poverty rate: from around one-half to over three-quarters of children are in the lower third.
Overall, we see in Figure 6 poverty rates in single-mother families of around 50 percent in the United States and slightly lower rates in Germany and Australia. In Canada and the United Kingdom the poverty rate for this family type is around 40 percent. Single-mother children's poverty rates are higher than two-parent rates in every country, of course, but the absolute differences are much larger in these five countries than in the others. We find a range of from around one-quarter to one-third in Spain, the Netherlands, France, and Italy. In the Nordic countries, Belgium, and Switzerland, single-mother child poverty ranges from 6 to 14 percent.
Figure 6. Percent of Children in Single-Mother Families with Low Income
In the United States we find an extreme poverty rate among single-mother families that is higher than the total poverty rate for this family type in nine other countries. We note that Italy and Germany have elevated rates of extreme poverty around 20 percent as well.
If there is another adult present in the single-mother family, the child poverty rate is generally much lower. On average the poverty rate with more than one adult present is less than half that when there is just the mother. But the United States is an exception. Even with additional adults, the poverty rate of children in single-mother families is close to 50 percent.
We noted that the poverty rate of children in two-parent families in the United States is 11 percent. We find now that in two countries the poverty rate of children in single-mother families with no other adults is in fact lower—in Finland it is 8 percent and in Sweden 5 percent. In Belgium it is the same as the U.S. two-parent rate. If one considers that two-parent families have two adults who may work, the comparison of the U.S. two-parent rate with that of single-mother families with two or more adults finds lower poverty rates in four additional countries—Denmark, Finland, the Netherlands, and Norway. Thus, we need to ask what it is about the income packaging in these countries that allows single mothers to escape poverty so much more frequently than even two-parent families in the United States.
Government and Market Income of Single Mother Families
In most of these countries a solid majority of children in single-mother families would be poor if their families received only income from the market. The proportion of pre-transfer poor ranges from just under 50 percent in Denmark and Finland to over 80 percent in the Netherlands and the United Kingdom (Figure 7). Switzerland, Spain, Belgium, and Sweden are in the 55 to 60 percent range of child poverty if we consider market income alone. Canada, Germany, Italy, France, Norway, and the United States cluster in the 62 to 66 percent range. Then there is a rather large jump to Australia, the Netherlands, and the United Kingdom, where pre-transfer child poverty rate is around 80 percent.
Figure 7. Pre- and Post-Transfer Child Poverty Rates in Single-Mother Families
There are very large differences in the proportion of pre-transfer poor moved out of poverty by transfer income. In the Nordic countries as well as in Belgium and Switzerland three-quarters or more of the pre-transfer poor children of single mothers are moved out of poverty by transfer income, with a high in Sweden of over 90 percent. In the Netherlands almost 70 percent of these children escape poverty via transfers. In Spain, France, and the United Kingdom a little over half of the pre-transfer poor children are moved out of poverty by transfers, slightly less than half in Italy, and some 40 percent in Australia. In Canada one-third of single-mothers' children are moved out of poverty by transfers. But in the United States and Germany only about one-quarter of single-mother poor children are moved out of poverty by transfers.
Thus, it is very clear that most single-mother families cannot depend on market income alone to rescue them from poverty or near poverty. Social transfers are an integral part of the income packaging of single mothers. Figure 8 highlights the role of market and transfer income in a different way. Here we plot the average market (earnings and asset) income and transfer income received in single-mother families in the lowest child income quintile. The figure charts the average income from market sources and the average income from transfer sources as a percentage of median equivalent income. The poverty rate for children of single mothers is given in parentheses after the country abbreviation.
Figure 8. Mean Factor and Transfer Income of Single-Mother Children in the Bottom Child Quintile
The five countries with the lowest poverty rates vary quite a bit in the mix of income types. Denmark has a high average of market income and a relatively low average of transfers; Sweden and Finland have lower average market income and higher average transfer income, followed by Norway; and at the extreme of reliance on transfer income we find Belgium. There is the same variation in the mix of income types among the countries with middling and high child poverty rates. Switzerland and Spain show a higher reliance on market income than France and, in particular, the Netherlands, both of which rely on transfers to a much greater extent. Yet all four countries have somewhat similar poverty rates. Finally, among the countries with very high poverty rates, the United States and Germany rely most heavily on market income while the United Kingdom and Australia rely on transfers. Thus, countries may differ a great deal in the content of the income package and yet produce similar poverty rates.
We find an extremely wide range across countries in the proportion of low-income children with no market income at all. The leader here is the United Kingdom, where 85 percent of lowest-quintile single-mother children live in families that depend completely on transfer income. Belgium, Australia, and the Netherlands come next with almost two-thirds of single-mother children in such families. In Germany and Canada we find about half of these children living without market income. Italy, the United States, Finland, France, Norway, and Spain range from 40 percent downward to one-third of single-mother children without market income. And, as noted above, Denmark, Sweden, and Switzerland have few children in families with no market income—ranging from 28 to 18 percent.
Low-Income Single Mothers' Labor Force Participation and Earnings
Almost all the market income for low-income single-mother families comes from earnings. In most cases, as we have seen, there is only one adult in the household and therefore only one person who can have earnings. For the most part, then, the variations we see in the role of market income are simply variations in the earnings of the single mother who heads the household. We consider here the labor force participation of single mothers whose children are in the lower third of the equivalent income distribution of all persons.
The fact that in the four Nordic countries around three-quarters of these children's single mothers have earnings is very much a product of jobs associated with the growth of the social service sector in these countries. In Denmark and Sweden around two-thirds of single mothers with earnings in this low-income group have jobs in the government/social service sector. In Finland the proportion of single mothers with government or social service jobs is smaller but nevertheless fairly high—around half of those with earnings have jobs in the government sector. (Unfortunately we do not have industry data for Norway.) The jobs these women hold tend to be less skilled jobs, such as caregivers in day-care centers for both children and the elderly, in schools, and in medical settings. But although these are not generally professional jobs, they do provide a certain amount of security, with full fringe benefits and standard wages.
Belgium has a somewhat similar pattern in terms of government versus private sector jobs, although the overall proportion of mothers with earnings (40 percent) is quite a bit lower than in the Nordic countries. We do find that the majority of those with earnings in Belgium have jobs in the government sector. In France a majority of those with earnings are in the government sector also, but the proportion of low-income single mothers with earnings (about two-thirds) is a bit lower than in the Nordic countries.
A very different pattern is found in the United States, the Netherlands, and the United Kingdom. Despite the fact that two-thirds of single mothers in the United States have earnings, fewer than one-fifth of them are found working in the government sector. Instead they are exposed to the low wages and marginal security of the least desirable jobs in the private sector. In the case of the Netherlands and the United Kingdom, the proportion with government jobs is quite small because the proportion with a job at all is small (less than 30 percent).
When the mother has earnings, we find a similar variation across countries in how much she earns. But there is only a weak positive association between the likelihood of a single mother having earnings and the average amount she earns. The working low-income mothers with the highest earnings are in Denmark, Finland, Sweden, Switzerland, Germany, and Belgium—averaging from 40 percent of median income to a high of 60 percent in Denmark. The low-earning single mothers (30 percent or less) are in Australia, the United Kingdom, and the United States.
In sum, if we think of the poverty rate for children in single-mother families as a function of their mothers' earnings and social transfers, we find that across these 15 countries, market income (principally earnings) seems to play a larger role than transfers, although both are important. The correlation of the single-mother child poverty rate with low-income single mothers' mean market income is -0.70. Its correlation with mean transfers is much lower (-0.25).
Conclusion: Finding the Will Is Finding the Way
This paper's findings underline the need for a comprehensive policy to reduce U.S. child poverty rates and improve the well-being of children. Human capital concerns in industrialized countries offer an overwhelming and very practical case for adequate investment in succeeding generations. Especially in the affluent economies of the industrialized world, there are no valid economic excuses for high child poverty rates. The human costs of poverty are high for parents and children alike, as is amply documented by such recent studies as Making Ends Meet: How Single Mothers Survive Welfare and Low-Wage Work by Kathryn Edin and Laura Lein (1997) and On (Not) Getting By in America by Barbara Ehrenreich (2001). These studies show us the full complexity of living a poor life- its stresses, dangers, few pleasures, and ultimately stunting effects. We can all afford low child poverty rates.
No two countries fight poverty in exactly the same way. Each nation's policies must fit its national culture and values. Thus policy suggestions must be hewn together into a system of child poverty reduction that will work in the United States. Moreover, these policies cannot simply be taken off the shelf and plugged in. It takes political leadership to make such polices national priorities and to make programs mesh in a supportive fashion.
The U.S. social system has now achieved a primary goal: reducing welfare dependence. However, as welfare rolls have been trimmed, corresponding decreases in child poverty have not been achieved. To experience serious reductions in child poverty, the United States needs to make this goal a top priority for its political agenda. The integrity of our democratic values will be assured and the cultural and economic fabric of our society enriched when we can say that no children grow up poor in America.
Endnotes
1. Rainwater, Lee and Timothy M. Smeeding, pages 1 and 2, from “Introduction.” Pages 17-22 from “Child Poverty Rates in Rich Countries in the 1990s: An Overview.” Pages 32-39 from “Patterns of Child Economic Well-Being.” Pages 68-76 from “Income Packaging: Market in Income and the State.” Pages 79-89 from “Child Poverty and Income Packaging in Two-Parent Families.” Pages 109-122 from “Child Poverty and Income Packaging in Single-Mother Families.” Page 141 from “Is There Hope for America's Low-Income Children?” In Poor Kids in a Rich Country. ©2003 Russell Sage Foundation, 112 East 64th Street, New York, NY 10021. Reprinted with permission.
2. Up-to-date information on the projects, countries, and variables included and the working papers published can be found at www.lisproject.org.
3. A reader might be puzzled to find here no discussion of the relation of patterns of race and ethnicity to poverty and income distribution in the countries studied. Clearly ethnicity is a strong correlate of economic disadvantage in the United States and in some other countries. But, in a cross-national context its status as a “cause” is ambiguous.
It is difficult to define ethnicity across countries in any way that one might call comparable. No other country has the history of slavery, de jure segregation, and continuing discrimination that one finds in the United States. The countries in our study have much smaller minority populations than the United States – on the order of ten percent or less compared to one-third. Each country has a different history of how it counts minorities so that there are many pitfalls to trying to define a group that is comparable to a U.S. “race” or “ethnic group.” For example, France categorizes by citizenship so that people of North African heritage who were born in France or are naturalized cannot be considered separately from those of French background. And, finally, some of the LIS surveys used do not provide any information on nationality. Our very rough conclusion from what data is available is that while in most (or perhaps all) countries recent arrivals have elevated child poverty rates compared to natives these difference nowhere approach those in the United States between minorities and whites. And, it is not clear that the differences that do exist are likely to persist over the decades in the same way they have in the United States.
4. We deviate from common definitions here by not including private transfers—primarily child support—in pre-transfer income. We do so because private transfers often are mandated by and payment enforced by the legal system so we categorize them as post government.
5. A reminder: cohabiting couples are included in two- parent families. Single parents do not have a cohabiting partner.
Timothy M. Smeeding is the Maxwell Professor of Public Policy, and Director of Maxwell's Center for Policy Research. He is also the Director of the Luxembourg Income Study (www.lisproject.org), a nonprofit research organization which he co-founded with Lee Rainwater in 1983. Professor Smeeding's research is focused on national and cross-national aspects of economic inequality, poverty, and public policy toward vulnerable groups, such as children, the aged, and the disabled. Over the next eight months he will publish three books: The Economics of an Aging Society (with Richard V. Burkhauser, Robert L. Clark, Marilyn Moon, and Joseph F. Quinn; Blackwell Publishers); Poor Kids in a Rich Nation (with Lee Rainwater; Russell Sage Foundation Press); and Public Policy and the Future of the Family, proceedings of a conference held at The Maxwell School in October 2002 (edited with Lee Rainwater, and Professor Daniel Patrick Moynihan; Russell Sage Foundation Press). Additional unpublished papers and a complete curriculum vitae may be found on his website at http://www-cpr.maxwell.syr.edu/faculty/smeeding/
Lee Rainwater is professor of sociology emeritus and research director of the Luxembourg Income Study. His books include: Poor Kids in a Rich Country: America's Children in Comparative Perspective, with Timothy M. Smeeding, Russell Sage Foundation, 2003; Income Distribution in OECD Countries, with Anthony B. Atkinson and Timothy M. Smeeding, OECD, 1995; Public/Private Interplay in Social Protection with Martin Rein, (eds.), M.E. Sharpe and Co., 1986; What Money Buys: Inequality and the Social Meaning of Income, New York: Basic Books, 1974; Behind Ghetto Walls: Black Families in a Federal Slum, Chicago: Aldine Publishing Company, 1970; and And the Poor Get Children: Sex, Contraception and Family Planning in the Working Class, Chicago: Quadrangle Books, Inc., 1960.
References
Citro, Constance F. and Robert T. Michael (1995). Measuring Poverty: A New Approach. Washington, DC: National Academy Press. http://books.nap.edu/books/0309051282/ html/index.html
Duesenberry, James S. (1949). Income, Saving and the Theory of Consumer Behavior. Cambridge: Harvard University Press.
Edin, Kathryn and Laura Lein (1997). Making Ends Meet: How Single Mothers Survive Welfare and Low-Wage Work . New York: Russell Sage Foundation.
Ehrenreich, Barbara (2001). Nickel and Dimed: On (Not) Getting By in America. New York: Metropolitan Books.
European Commission (1985). “On Specific Community Action to Combat Poverty (Council Decision of 19 December 1984).” Official Journal of the EEC 2/24 (85/8/EEC).
Harrington, Michael (1962). The Other America: Poverty in the United States. New York: Macmillan.
Johnson, Lyndon B. (1964). “Proposal for a Nationwide War on the Sources of Poverty: Lyndon B. Johnson's Special Message to Congress, March 16, 1964.” Quoted with permission from Public Papers of U.S. Presidents, Lyndon B. Johnson, 1963-1964 (Washington: G.P.O., 1965), 1, pp. 375-380. Created by Internet Modern History Sourcebook. Last updated 1988. http://www.fordham.edu/halsall/mod/1964johnson-warpoverty.html
Katz, Michael B. (1986). In the Shadow of the Poorhouse: A Social History of Welfare in America. New York: Basic Books.
Katz, Michael B. (1989). The Undeserving Poor: From the War on Poverty to the War on Welfare. New York: Pantheon Books.
Lancaster, Kelvin (1971). Consumer Demand: A New Approach. New York: Columbia University Press.
Moynihan, Daniel Patrick (1996). “Opinion: When Principle Is at Issue.” The Washington Post . Sunday, August 4, C07.
Patterson, James T. (1981). America's Struggle against Poverty, 1900-1980. Cambridge: Harvard University Press.
Peattie, Lisa, and Martin Rein (1983). Women's Claims. Oxford: Oxford University Press.
Rainwater, Lee, Martin Rein, and Joseph E. Schwartz (1986). Income Packaging in the Welfare State. Oxford: Oxford University Press.
Rainwater, Lee, and Timothy M. Smeeding (2003). Poor Kids in a Rich Country. New York: Russell Sage Foundation.
Ruggles, Patricia (1990). Drawing the Line: Alternative Poverty Measures and Their Implications for Public Policy. Washington, DC: Urban Institute Press.
Sen, Amartya Kumar (1992). Inequality Reexamined. New York/Cambridge, MA: Russell Sage Foundation/Harvard University Press.
Smeeding, Timothy M., Lee Rainwater, and Gary Burtless (2001). “United States Poverty in a Cross-National Context.” In Sheldon H. Danziger and Robert H. Haveman, eds. Understanding Poverty. New York and Cambridge, MA: Russell Sage Foundation and Harvard University Press, 162-189.
Vaughan, Denton R. (1993). “Exploring the Use of the Public's Views to Set Income Poverty Thresholds and Adjust Them over Time.” Social Security Bulletin 56(2): 22-46.