Alternatives II: Social Action

excerpted from the book

Crime and Punishment in America

Why solutions to America's most stubborn social crisis have not worked - and what will

by Elliott Currie

Henry Holt, 1998, paper


... there is now overwhelming evidence that inequality, extreme poverty, and social exclusion matter profoundly in shaping a society's experience of violent crime. And they matter, in good part, precisely because of their impact on the close-in institutions of family and community. We do not know all that we need to know about these connections, but a consistent picture emerges from the accumulated research. And it is a picture that most of us, on reflection, would not find very surprising. On the face of it, the argument that social factors are unimportant in explaining crime is an improbable one. Most people, after all, understand viscerally that some places are more dangerous than others and that this has to do, in part, with the social conditions within them. If they travel, they know that some cities are more dangerous than others; that walking the streets of Washington, D.C., at night is not the same sort of experience as walking the streets of Copenhagen, or even Toronto. At home, most Americans would prefer that their car break down, if it must, in a leafy middle-class suburb than in the heart of the urban ghetto.

Nevertheless, a small but influential handful of social scientists and pundits still insist that social and economic differences have little or nothing to do with crime rates-and that, accordingly, expanding opportunities and providing better supports for the vulnerable won't make any difference.

Around the world, the countries with relatively low levels of violent crime tend to be not only among the most prosperous but also those where prosperity has become most general, most evenly distributed throughout the population. The countries where violent crime is an endemic problem are those in which prosperity, to the extent that it is achieved at all, is confined to some sectors of the population and denied to others. That includes a number of less developed countries in Latin America, Africa, and the Caribbean (and parts of the former Soviet bloc) and one country in the developed world-the United States.

What Americans have difficulty comprehending is that in more advanced countries, death by deliberate violence is a very rare phenomenon-sufficiently rare that it does not appear in statistics of the leading causes of death. In the United States, homicide is the third leading cause of death for black men of all ages, fifth for Hispanic men, and seventh for Asian men. Astonishingly, homicide even ranks among the top ten causes of death for both black and Asian women. The composition of deadly violence is different as well. Homicide in the United States is much more likely to involve strangers than it is in other countries. In Sweden, for example, homicide is not only low by our standards-averaging considerably less than 2 per 100,000 per year for decades (versus our 8 to 10 per 100,000)-but has historically occurred "overwhelmingly between relatives and persons known to each other." The United States, in short, as people in less violent countries understand and routinely comment upon, is a strange and disturbing anomaly in the developed world.

These differences are not confined to homicide, although it's there that our statistics are most reliable and, probably not coincidentally, that the international differences are sharpest. Take robbery: although even minor robberies are apparently more likely to be reported and recorded in many other developed countries (because their citizens are less resigned to being robbed and less fearful of reporting crime to the police), reported robbery rates are much higher here than in comparable countries. Thus, despite the several-year decline in crime in New York City, it was only in 1995 that the number of robberies there fell below that in all of England and Wales, with seven times the population.

Robberies in America, moreover, are far more likely to be committed with weapons. And it is the combination of the greater likelihood of violent crime in the first place with the higher probability that the crime will involve a weapon, especially a gun, that accounts for the uniquely high levels of deadly violence in the United States. Franklin Zimring and Gordon Hawkins have detailed this pattern in comparing robberies and burglaries in London and New York City. New Yorkers are far more likely to be robbed (the rate of reported robberies in 1992 was nearly four times that in London); and their chances of being killed during a robbery are even more disproportionate. There were 5 robbery deaths in London in 1992, versus 357 in New York. That difference, not surprisingly, is closely tied to the far more frequent presence of guns in New York's robberies. Startlingly, however, the risk of being killed in the course Of a robbery in New York that does not involve a gun was still four times greater than the risk of being killed in any robbery in London.

Compared with the British, then, the American pattern is distinctive at every level: Americans are far more likely to commit robbery in the first place, more likely to use a gun and thus risk the death or serious injury of their victim if they do, and more likely to inflict deadly force on their victim whether or not they use a gun. Among other things, this suggests that reducing the use of guns in predatory crimes could go a very long way toward reducing the worst American violence-an issue to which we'll return. But it also reminds us that the distinctive American levels of violence involve forces that go much deeper than the availability of guns alone.

It is hard, then, to credit the idea that "prosperity" is to blame for a problem which remains relatively tractable in most prosperous societies-with one notable exception. Blaming "freedom' for crime's ravages runs up against the same logical problem. Holland may be the "freest" country in the world, when it comes to tolerance of doing one's own thing (that has certainly been true of its attitude toward drug use). But Holland's murder rate, at last count, was one-sixth the U.S. rate.

... the stark differences in serious criminal violence among the wealthy industrial democracies make it clear that neither democracy nor freedom nor prosperity can explain these patterns-and that something that distinguishes the United States from most other industrial countries must be an important part of that explanation.


If prosperity, democracy, and freedom are not to blame for America's violence problem, what is?

The accumulating research suggests that the answers are complex. But there are answers. There are indeed "root causes" of America's disturbingly high levels of violence, and they involve precisely those things that have made America's brand of prosperity very different from that of other advanced societies.

We can begin to trace some of these differences in the findings of the Luxemburg Income Study (LIS), an international survey of poverty, inequality, and government spending in industrial countries, directed by Lee Rainwater of Harvard University and Timothy Smeeding of Syracuse University. With carefully assembled data and in cautious, understated language, the LIS paints a troubling portrait of the United States-a country that, though generally quite wealthy, is also far more unequal and far less committed to including the vulnerable into a common level of social life than are other developed nations.

To begin with, children and families in the United States are far more likely to be poor than are their counterparts in other industrial democracies-and, if they are poor, are more likely to be extremely poor. Our overall child poverty rate in the early 1990s was about 22 percent; Australia and Canada, our closest "competitors" in the industrial world, had rates of about 14 percent. Most of Europe was far behind, with France and Germany at around 6 percent and Sweden, Norway, and Denmark, among others, below 4 percent.

If These calculations are based on a relative measure of poverty, in which being poor is defined as having less than half the country's overall median income. But the LIS also shows that America's poor are poorer in absolute terms than their counterparts in almost every industrial nation in the world. It is often argued that poor people in America are really not so badly off at all-indeed, much better off than middle-class people in many other countries-proof that "economic poverty" cannot be a significant cause of crime or other American social ills. But the LIS shows that the reality is quite different. Rich American children are indeed considerably better off, on average, than rich children in most other advanced societies; middle-income American children also tend to be somewhat better off than middle-income children elsewhere. But poor American children are considerably worse off- sometimes dramatically so-than their counterparts in other advanced countries. Measured by the real income available to their families, Swiss and Swedish children in the bottom 20 percent of the income distribution have 72 percent more income than their American counterparts; German children, 40 percent more; Canadian children, 25 percent more. Of the eighteen countries studied by the LIS, only in Ireland and Israel-two relatively low-income countries-are poor children poorer, in standardized dollar terms, than their American counterparts.

Accordingly, the economic condition of children in America combines both absolute deprivation and inequality; our poor children are poorer and our rich children are richer than those in comparable societies. And these figures actually understate the relative plight of low-income American children because they are based on what economists call "purchasing-power parities" (that is, on what families can buy in the private market). They mask the fact that in other advanced nations certain basic social services-including health care and child care-are often provided by the government as a matter of right. Factoring in the publicly funded national health care in Britain or Ireland, for example, would raise the estimate of poor children's real income in those countries, reduce the measured inequality among their children, and widen the gap between the economic well-being of poor children there and in the United States.

We know-from both the LIS and other studies-that there are two key reasons for these differences. The first is that Americans who work in the lower reaches of the paid labor market are more likely to earn very low wages than their counterparts elsewhere. The problem of the "working poor," in other words, is much more severe in the United States. The Harvard economist Richard B. Freeman points out that American workers in the lowest 10 percent of the pay scale earn, on average, just 38 percent of the national median wage for all workers-versus 61 percent among Japanese and 68 percent among German workers. The result is that the United States maintains a very high rate of what economists refer to as "pretransfer" poverty-poverty before government steps in with public income supports to buffer the effects of low (or no) earnings.

But the LIS makes startlingly clear that the United States also does less to offset the problem of poor earnings through government benefits. The United Kingdom, for example, also has a high rate of family poverty before government transfers, but a moderate (though recently increasing) rate after the government steps in with income assistance (again, this is true even without factoring in the value of free health care under the British national health system). In Germany and Scandinavia, generous government income supports for families and children come on top of already high incomes from work, which makes for exceptionally low rates of "posttransfer" poverty. In Sweden, the rate of posttransfer poverty among single-parent families is less than half the American rate for two-parent families. This may change, as both the Swedish commitment to full employment at high wages and their generous system of social provision have come under attack in the name of fiscal austerity. But historically, these two factors have combined to keep poverty in Sweden among the lowest in the world.

A similar picture of America's distinctive pattern of high inequality and low social support emerges if we calculate how much various countries spend on what Europeans call "social protection as a proportion of their overall economic wealth. Thus, counting not only what we normally think of as "welfare" but also family allowances, unemployment compensation, and disability benefits, the United States spends less than 4 percent of its Gross Domestic Product (GDP) on Such programs, and the proportion has fallen steadily since 1980. At the other extreme, the Scandinavian countries and the Netherlands spend from 12 to more than 14 percent, and the proportion has generally risen in recent years-sharply in some countries. Even after years of conservative rule, Britain spends twice what the United States does, proportionately, on social protection-as does Ireland, again a relatively poor but, within its economic constraints, quite generous society by American standards.

The bottom line, then, is that American children and their families are forced to make do in the vicissitudes of a volatile market economy with unusually little public support. They are far more likely to suffer a kind and degree of social exclusion that is no longer tolerated in other industrial societies-even ones that in sheer material terms are considerably poorer than the United States. It's important to be clear that this is not just a difference in material conditions but also a difference of values-of culture. Our willingness to tolerate extremes of deprivation and social insecurity points to deep cross-national differences in the most basic conceptions of collective responsibility for the well-being of others. Among the industrial societies at the close of the twentieth century, the United States stands out as uniquely Darwinian, a prime example of what the British criminologist David Downes calls a "winner-loser culture."

And these differences are becoming more pronounced, for both inequality and poverty have recently increased more rapidly in the United States than in most other developed countries. The numbers are familiar, and we need only touch on a few that reveal the magnitude of these shifts in the United States. The early 1970s as we've seen, marked the beginning of the American prison boom they were also the low point in American rates of poverty and economic inequality. About 11 percent of families with children under eighteen were poor in 1973; the figure reached almost 19 percent by 1993 and, despite a strong economic upswing, was still over 16 percent in 1995. And a greater proportion were not just poor but desperately poor. In 1976, 28 percent of poor children lived in families earning less than half of the federal poverty line; by 1995, 39 percent did. Among black children, the proportion in severe poverty doubled in those years: nearly half of poor black children in 1995 (in the midst of an economic boom) were living below 50 percent of the federal poverty level. Partly because of the spread and intensification of poverty, the economic gap between haves and have-nots in the United States increased as well. In the mid-1970s, families in the upper 20 percent of the income scale I had about 7 1/2 times the income of those in the bottom 20 percent by the mid-1990s, 11 times.


Countries where there is a wide gap between rich and poor routinely show higher levels of violent crime-which helps explain why the world's worst levels of violence have been found in places like Colombia, Venezuela, South Africa, and Mexico, where inequalities are even harsher and more consequential than in the United States. Look closer, and it becomes apparent that violence is worse in neglectful or mean-spirited societies than in more generous ones-even if they are poorer. Societies with weak "safety nets" for the poor and economically insecure are more likely than others at a comparable level of development to be wracked by violence.

That is one conclusion, for example, of a study by Rosemary Gartner of the University of Toronto, who examined homicide rates in eighteen developed countries from 1950 to 1980. Even in the earlier part of the period, the United States already suffered considerably higher rates of homicide than the other countries, and the composition of American homicide was different as well: Americans homicide victims were more likely to have been murdered by strangers, as opposed to intimates, than were murder victims in other countries. And though homicide increased somewhat for the eighteen countries as a whole over the thirty-year period, the extraordinary dominance of the United States remained unchanged. On average, American men died of homicide at an annual rate of about 14 per 100,000 during these years. The next-highest rate-about 4 per 100,000-was in Finland. No other country among the eighteen had a rate as high as 3 per 100,000, and several-including Denmark, England and Wales, Ireland, Holland, and Switzerland-had rates below 1 per 100,000.

What accounted for these differences? Gartner's study points to several factors. Economic inequality had a powerful effect on the countries' homicide levels. A measure of "social security" expenditure as a proportion of GNP-including cash benefits for social welfare and family allowances, along with unemployment insurance, public health spending, and other ameliorative programs-likewise had a strong effect on the risks of homicide for every age group. High divorce rates, ethnic and cultural heterogeneity, and a cultural leaning toward violence generally (as measured by support for the death penalty and frequent wars) also seemed to promote homicide. But the effects of both inequality and the relative absence of social provision remained powerful even when all else was accounted for.


The links between extreme deprivation, delinquency, and violence then, are strong, consistent, and compelling. There is little question that growing up in extreme poverty exerts powerful pressures toward crime. The fact that those pressures are overcome by some individuals is testimony to human strength and resiliency, but does not diminish the importance of the link between social exclusion and violence. The effects are compounded by the absence of public supports to buffer economic insecurity and deprivation, and they are even more potent when racial subordination is added to the mix. And this-rather than "prosperity"-helps us begin to understand why the United States suffers more serious violent crime than other industrial democracies, and why violence has remained stubbornly high in the face of our unprecedented efforts at repressive control.

... those countries with the most developed welfare states have far less violence than the United States, the industrial nation with the least developed welfare state.

The cross-national research conducted by Rosemary Gartner and her colleagues sheds more light on this issue. In one study Gartner and Fred C. Pampel of the University of Colorado explored the implications of a peculiar difference in the pattern of youth violence between the United States and other high-income countries. One of the most consistent findings about the homicide rate in this country is that it tends to be higher when there is a higher proportion of youths in the population. (A rising proportion of young people is one key reason, for example, why homicide rates rose sharply in the United States during late 1960s.) Some criminologists have taken this to mean that there is an ironclad connection between youth and violence. But when we look overseas, the connection evaporates. In many other industrial democracies, the proportion of youths in the population has little or no effect on the homicide rate. This suggests that it is not simply the fact of being young, but something about being young in the American social context specifically, that leads to higher homicide rates when the youth population grows. But what is that "something"?

Rejecting what they call a "naive demographic determinism," Gartner and Pampel found that what matters more than the sheer number of youths are the social protections a society provides for young people and their families. In societies sharing what they call a "collectivist" orientation-societies that, among other things, provide universal and relatively generous social benefits (including health care, income support, disability, unemployment insurance, and family allowances)-there is only a weak connection between the size of the youth population and the rate of homicide. In Norway, for example, a classic "social democracy," changes in the age structure have no effect on the level of homicide; in the U.S., an equally classic example of an "individualistic" political culture, rises in the proportion of youths in the population translate directly into higher homicide rates. The research is complex and sophisticated, but the basic message is simple: countries that have made a long-standing commitment to provide at least a modest floor of income and social inclusion for all their citizens have less youth violence than those in which children and families are routinely left to fall through the holes of an already threadbare "safety net."

Similarly, the few studies that look specifically at the impact of welfare benefits on rates of crime within the United States show that welfare actually tends to lower crime rates, not raise them. The University of Connecticut sociologist James DeFronzo, for example, has found that higher welfare payments appear to reduce rates of burglary. DeFronzo looked at Aid to Families with Dependent Children (AFDC) benefits per person, adjusted for differences in the cost of living, in 140 metropolitan areas across the United States. He also included, as "control" variables, the cities' level of poverty and unemployment, the number of households headed by women, the percentage that was black or Hispanic, the proportion of young males in the population, the per capita income, and the city population. With all of these other factors controlled, AFDC payments had a strong negative relationship to burglary rates-that is, the higher the average welfare benefit, the lower burglary rate.

Why might higher welfare payments reduce property crimes? DeFronzo's data cannot provide an answer, but there are several possibilities. One is that where poor families get only a bare minimum income from either work or welfare, illicit ways of earning income-whether from stealing, dealing drugs, or prostitution- will obviously be more attractive. Another, more complex, explanation may be that extremely poor families are less able to nurture and supervise children and teenagers in ways that keep them from committing this kind of crime. (Burglary is typically a crime of the young; the rates peak at about age fifteen.) Whatever the specific mechanism-and there may be several-the overall implication is that, as DeFronzo says, "lowering welfare support levels for the poor might increase crime rates." In a more recent study, DeFronzo has found a similar pattern for homicide; once other social and economic factors are controlled, high homicide rates are strongly correlated with low AFDC payments.


The key to our unique problem with violence, in short, is not the overdevelopment of the welfare state but the opposite: the underdevelopment of the mechanisms of social and economic inclusion that have blunted the edges of market capitalism in other industrial democracies.

But what is it about extreme social exclusion of the kind so prevalent in the "sink or swim" culture of the United States that breeds violence and delinquency with such distressing predictability? Knowing that there is a powerful connection between deprivation and crime doesn't by itself tell us exactly how that connection operates. Here again, we don't know all that we need to know, but some of the mechanisms seem increasingly clear from the accumulating evidence. Most of them are, on reflection, not very surprising. And they help us mightily to think about crafting solutions.

To give the skeptics their due, it's true that the connections between poverty, inequality, and violence aren't always either simple or direct. It isn't the lack of money alone that breeds violence; if that were the case, then graduate students would be very dangerous people indeed. It's rather that the experience of life year in, year out at the bottom of a harsh, depriving, and excluding social system wears away at the psychological and communal conditions that sustain healthy human development. It stunts children's intellectual and emotional growth, undercuts parents' ability to raise children caringly and effectively, increases the risks of child abuse and neglect, and diminishes the capacity of adults to supervise the young. It creates neighborhoods that are both dangerous and bereft of legitimate opportunities and role models, makes forming and maintaining families more difficult, and makes illicit activities far more alluring for teenagers and adults. Life, in short, is harder, bleaker, less supportive, and more volatile at the bottom- especially when the bottom is as far down as it is in the United States. And those conditions, in ways both direct and indirect, both obvious and subtle, breed violent crime.

In Europe, care for children aged three to five is virtually universal. Sometimes, as in Sweden, it is offered in freestanding child-care centers; more often, it is provided through heavily subsidized preschool programs connected with the school systems. Again, exactly how we provide a similar level of care is less important than that we do it one way or another. For older children, making school facilities widely available after regular school hours for learning and recreation (what some advocates call "second-shift" schooling) promises similar benefits.

The fact that we are the only advanced nation without a national system of subsidized health care also contributes to our crime problem in several ways.

It is now fashionable to argue that private philanthropy and voluntarism should take up the 1 tasks we are increasingly stripping from government. But though exhorting more Americans to volunteer to help the casualties of an increasingly unequal and insecure economy may be better than no response at all, it cannot compensate for the damage done by shortsighted social policies.

The present drift of national policy, indeed, reflects a stunning degree of collective denial. Across the political spectrum, for example, there is agreement that families are critical-that what happens in them has an enormous influence on whether our children grow up compassionate or predatory, moral or irresponsible. Yet as a society we want to have it both ways. We want competent, caring families that can do a good job of socializing and supervising children, but we refuse to provide the social supports that would make that possible. We force many parents to choose between draining overwork in the low-wage economy and demoralizing poverty outside it. We want parents to work, but we balk at providing the child care that would allow them to do so without jeopardizing the well-being of their children. We want parents to spend quality time with their children, but we reject the paid work leaves or shorter hours that would make it possible. Then we blame families for the consequences. We understand that childhood traumas may lead to violence, but we draw the line at reliably providing the preventive health care that could address them. We acknowledge the link between child abuse and violent crime, but we starve our child-protective systems. In the future, we will have to choose between perpetuating the kinds of family stresses that we know breed violence and finally bringing our family policies into line with those of most other industrial democracies-which, not coincidentally, suffer far less violent crime.

The same head-in-the-sand denial runs through our attitude toward work. Here, too, we seem to want contradictory things. We want people to work, and, of course, we want them to work in the legitimate labor market. But we also deliberately maintain unnaturally high levels of joblessness, partly because we are afraid that full employment will raise wages-and hurt the stock and bond markets. We support, rhetorically, the idea that everyone ought to be a productive member of the community. But in practice, we are apprehensive at the prospect of an economy in which everyone is engaged in productive work. That is the main reason why we haven't made a serious commitment to public job creation, even though such a commitment has been urged repeatedly since the Second World War. But we simply cannot have it both ways. We cannot insist on maintaining Depression-era levels of joblessness among the young and poor in order to preserve the income of affluent bondholders and simultaneously expect to keep our streets and homes safe from violent crime.

Crime and Punishment in America

Index of Website

Home Page