Ending Poverty in America: Using Carrots and Sticks

In the last decade, we have seen that an effective approach to reducing poverty requires changes in personal behavior as well as government support. Further, we have learned that by judiciously applying policies that demand and then reward good behavior–what might be called carrots-and-sticks policies–we can in-duce and maintain the behavior that leads to reduced poverty. Reviewing the record of the past decade sug-gests the principles that should guide future efforts.

During the 1960s, child poverty fell by more than half, to 14 percent. In the subsequent three decades, however, child poverty drifted upward in an uneven pattern, never again reaching the low level achieved in 1969. This is a surprising and discouraging record.

A major cause of the huge decline in poverty of the 1960s was an economy that grew 35 percent in per capita gross domestic product, giving rise to President Kennedy’s famous observation that “a rising tide lifts all boats.” Although the American economy has grown at a more stately pace since the 1960s, each subse-quent decade has nonetheless seen substantial growth of more than 20 percent in per capita GDP–plenty of raw material, one would think, to continue the poverty reduction that distinguished the ’60s. But the 24 per-cent GDP growth of the 1970s saw poverty fall by a mere 5 percent, and the 23 percent GDP growth of the ’80s saw poverty actually increase by 12 percent. Clearly, a rising tide was not lifting all boats. The second half of the ’90s, however, once again saw strong economic growth accompanied by the fastest and deepest decline in child poverty since the ’60s.


Three trends tell us a lot about the causes of poverty and show us why a growing economy has not been more effective in reducing it. First, as the left chart shows, growth of wages at the bottom of the distribution (the 10th percentile) declined during the 1980s and the first half of the ’90s, rising again only after 1996.

High Employment. Stagnant or falling wages at the bottom of the distribution make reducing poverty dif-ficult. By contrast, tight labor markets, as signaled by low unemployment rates, contribute to both rising wages and falling poverty rates.

Consider the record: Wages rose and poverty fell during the 1960s, when unemployment averaged 4.8 percent and fell as low as 3.5 percent. But as wages fell or were stagnant during the ’70s and ’80s, when unemployment skyrocketed to average 6.2 percent and 7.3 percent, respectively, poverty rose or was stag-nant. Only when tight labor markets returned after the mid-’90s–when unemployment fell to an average of 4.8 percent between 1995 and 2000–did wages once again rise and poverty fall. Mere economic growth will not necessarily lead to reduced poverty rates. Apparently, tight labor markets accompanied by rising wages are required to effectively fight poverty.

Family Factors. A second factor putting substantial upward pressure on poverty was changes in family composition. The poverty rate for mother-headed families is usually four or five times the rate for married-couple families. So, other things being equal, any rise in the share of children living in female-headed fami-lies will increase poverty.

Beginning in the 1960s, Americans perfected every known method of casting children into single-parent families. Marriage rates fell, divorce rates increased until the 1980s, and non-marital birth rates exploded until a third of all babies (and nearly 70 percent of black babies) were born outside marriage. As a result, be-tween 1970 and 2004, the percentage of children living in a female-headed family increased from 12 percent to 28 percent. It’s hard to fight poverty when more and more children are in families of the type that are most likely to be poor.

Education. Poor educational achievement is a third reason poverty has been stagnant. Education has always been important in accounting for economic success, but most analysts agree that recent decades–because of globalization, technological change, and trade–have seen increased payoff to education. One of the most important changes in the American economy for those interested in fighting poverty is the decline of high-paying jobs suited to workers with a high-school education or less. Workers without a high-school di-ploma are twice as likely to be poor as those with one, and three times as likely to be poor as workers with some college education. The Educational Testing Service estimates that nearly one-third of students drop out of school before graduating. Moreover, despite waves of educational reform, the reading and math achievement of students from poor and low-income families has been virtually flat for three decades.

So there are at least three raging rivers against which those who would fight poverty must struggle: low wages, the rise of single-parent families, and lousy education. To offset these currents, the nation has spent an increasing amount of money on government programs to fight poverty. Between 1968 and 2004, the total of inflation-adjusted federal and state spending on means-tested programs (those that specify an income level above which individuals or families cannot qualify for benefits) increased from $89 billion to $585 billion–all without reducing poverty below its late-1960s level.

The net impact of these factors–economic growth, wages, family composition, education, and govern-ment spending–is high child-poverty rates. It is especially notable that the three factors over which individu-als have full or partial control–work, family composition, and education–were either stable or moved in the wrong direction. Until recently, millions of Americans failed to work, many languishing on welfare. Millions also decided to have children outside marriage, to avoid marriage, or to divorce. And millions of young peo-ple refused to apply themselves during their school years, eventually either dropping out of school or gradu-ating with low reading and math skills.


Yet the mid-1990s saw a dramatic example of how public policy can both help individuals improve their choices and reward them for doing so, namely the 1996 welfare-reform law, passed on a huge bipartisan vote in Congress and signed by Democratic President Bill Clinton. Perhaps the law’s single most notable fea-ture was that it made cash welfare contingent on individuals working or preparing for work. Individuals who did not meet work requirements had their cash benefit reduced, and in most states even terminated.

In addition, the law limited federal benefit receipt to five years for any given parent. The work require-ments, reduction of benefits for those who did not work, and five-year time limit all served as sticks that en-couraged or forced parents on welfare to work. Much to everyone’s surprise, mothers went to work by the hundreds of thousands, and the welfare rolls declined by more than 60 percent, far more than ever before. But most important, poverty among children in single-parent families fell by 30 percent, reaching its lowest level ever. Not surprisingly, given the high proportion of black children living in female-headed families, black child poverty also reached its lowest level ever.


Government-imposed work requirements were an important part of this welcome decline in child poverty. But work is only half the picture. Most welfare mothers, who typically can qualify for jobs paying about $8 per hour, were not able to earn enough money to bring their families out of poverty. Realizing that financial pay-offs to work were an important part of helping low-income families, federal and state governments, over a period of more than two decades, created and expanded programs specifically designed to help low-income working families. These included Medicaid health insurance, child care, food stamps, and above all, the Earned Income Tax Credit (EITC), a taxpayer-provided wage subsidy that could give working parents up to $4,000 in cash (in 1996).

Census Bureau data for children living with their single mothers, present a clear picture of why this new system of earnings from increased work effort supplemented by benefits from work-support programs led to such a dramatic reduction in child poverty. In the second chart, the bar graphs on the left are for 1990; those on the right are for 1999. The first graph in each set shows the poverty rate that characterized these families based only on their market income before any government payments–poverty in the state of nature, so to speak. Here it can be seen that increased work by mothers between 1990 and ’99 led to a huge 11 percent-age point reduction in poverty. When government cash and in-kind benefits other than those delivered through the tax code are added, poverty falls 13 percentage points in 1990. But poverty also falls by 9 per-centage points in 1999, demonstrating that working families are receiving work-support benefits to supple-ment their earnings to further reduce market poverty. Adding tax benefits, primarily the EITC, does not re-duce poverty at all in 1990, but reduces it another 5 percentage points in 1999. The combination of work and work supports reduces poverty a full 12 percentage points–or by about 4.5 million people–more in 1999 than in 1990. It was the stick of welfare reform that induced mothers to leave welfare for work; it was the carrot of work-support benefits that supplemented the mothers’ earnings and led to substantial reductions in poverty.

There are lots of good ideas for further reductions in poverty–improving and expanding preschool educa-tion; improving the public schools, especially for students from poor families; reducing nonmarital births; in-creasing marriage rates; encouraging savings; and helping poor young men improve their earnings–but the decline of poverty among female-headed families in the 1990s illustrates the principles that should guide the nation’s efforts. The first is that individuals must change their behavior–or the nation will not be able to sub-stantially reduce poverty. The second is that policy-makers should seek out policies that encourage or de-mand responsible behavior, and then use public dollars to reward it.