Lawrence M. Mead is associate professor of politics at New York University. He is the author of Beyond Entitlement. This article is based, in part, on his forthcoming book, The New Politics of Poverty (Basic Books).

New York is the capital of welfare, but not of welfare reform. No other city is so known for entrenched dependency. In 1988, nearly 700,000 city residents lived on Aid to Families with Dependent Children, the core welfare program. If other welfare programs are added, the total comes to more than 1.2 million, or 15 percent of the population, making the city’s welfare caseload larger than the total population of all but a handful of American cities. Mainly because of New York City’s lopsided figures, New York State also ranks near the top of all states for welfare dependency.

Nevertheless the city and state have virtually ignored the surprisingly successful wave of welfare reforms that swept other states and Washington in the last decade. Nearly all the recent welfare reforms have been based on what is now widely known as “workfare”: programs that require welfare recipients to work or prepare for work in return for support. Workfare used to mean requiring that employable recipients “work off,” or earn, their grants by taking unpaid government jobs. Today such programs may offer a range of activities, including private-sector job search and training, as well as mandatory unpaid work assignments.

Workfare raises family income, often reduces dependency, and most importantly, helps hard-core welfare recipients reorganize their lives. Workfare does not, for the most part, force people to do things they do not want to do. On the contrary, it helps welfare recipients do what they themselves wish to do: become more self-reliant.

Many welfare recipients live in a “culture of poverty” that lacks the practical sanctions, social pressures, role models, and other influences that uphold self-reliance for most Americans. As Congress recognized in the Family Support Act of 1988, which all but requires states to institute extensive workfare programs, workfare re-creates those sanctions, pressures, and supports.

In New York, however, bureaucracy, ideology, and politics have combined to frustrate workfare. Instead, the opponents of workfare advocate a range of voluntary approaches, including “work incentives” and training and jobs programs. Unfortunately, the evidence shows that hard-core welfare recipients seldom participate in such programs voluntarily. Few sign up, and many of those who do sign up do not show up. Under workfare, many more recipients sign up and show up, for work or for training, day after day.

This is a crucial difference. The nonworking poor are the reservoir from which our greatest social problems flow. The source of their difficulties, however, is not so much that they are dependent on a government check—so are Social Security recipients—but that they do not work, or come from families of workers. They are thus cut off from the social and economic discipline most crucial to integrating people into American life. Of course, to say that the problem of the nonworking poor is that they do not work is a truism. But it is a useful one, for it clarifies the path to reform: the way out of the welfare trap is not to solve all the subtle mysteries of dependency or refashion the economy, but to get the poor to work.

The past decade has seen some very impressive intellectual battles over the causes of and possible solutions to long-term dependency. And yet most of the discussion, even among thinkers who disagree, has been based on a common premise which we might call the “competence assumption.” This is the belief that hard-core nonworkers who live deep in the culture of poverty are nevertheless fundamentally like people in the rest of society. In particular, this view assumes that they respond to incentives and opportunities in essentially the same way as members of the middle class. Unfortunately, the long-term poor and dependent are seldom competent in this sense.

Conservatives, for instance, often argue that welfare per se causes dependency: Families become dysfunctional simply because welfare offers them an alternative to supporting themselves, an alternative which at least for a time can seem economically rational. There is some truth in this view: AFDC plus associated benefits such as food stamps, Medicaid, and public housing can compete in value with low-skilled wages, especially in places with high benefits such as New York. In this view, if we simply lowered benefits or limited eligibility, the poor, being as rational as anyone else, would leave welfare and get jobs.

But life on welfare seldom makes economic sense, even in the few cases where it may “pay” in the short run to be dependent. Federal figures show that most welfare recipients who take jobs soon find themselves above the poverty level and off the rolls. Steady work at any legal wage is a much surer way to lift a family above the poverty line than reliance on welfare, even in New York, and steady workers seldom remain at the minimum wage for long. Nationally in 1989, only 3 percent of family heads who worked full time all year were poor, according to Census Bureau data. The dependent also forgo the chances for advancement and higher wages available even to low-skilled workers. Moreover, the level of dependency turns out to be only loosely related to benefits. The states with the highest dependency rates in the country are Mississippi and Louisiana, where AFDC benefits are only $120 and $190 a month respectively. New York City’s AFDC benefits have declined by 37 percent in real terms since 1970, yet much of the increase in dependency occurred during that period.

Even most welfare recipients realize that welfare does not make sense: About half leave the rolls within two years after resolving the crisis that made them apply for help in the first place. The other half of the cases are much more worrisome, particularly the quarter that spend a total of ten years or more on AFDC. The key to welfare reform is to figure out why these people, the hard-core dependents, stay on the rolls instead of going to work.

More liberal analysts say that recipients are not tempted onto welfare by the benefits but forced onto it by injustice, economic adversity, or a lack of social support. They contend that the poor and dependent cannot work because the wages they could earn are too low, jobs are unavailable due to racial bias or the collapse of manufacturing employment, or welfare mothers are hamstrung by children or an absence of child care. But research has undermined each of these claims.

It is true that low-skilled workers are often poorly paid, but as noted above, the rewards of employment, if sustained, almost always exceed those of welfare. On the whole, it is not low-skilled workers who suffer racial exclusion but more affluent blacks seeking positions traditionally held by whites.

It is also implausible to say welfare mothers are prevented from working by children. Nationally, more than half of single mothers are employed, and nearly three of four employed single mothers work full time. Welfare families are not even unusually large; three-fourths of the mothers have only one or two children, according to Congressional figures. Child care is vastly more available than welfare advocates claim. According to government surveys, most child care is cheap arranged informally with friends or relatives. Government child care centers are secondary.

Nor are jobs scarce. The decline of the factory has been offset by an explosion of jobs in the service economy. Immigrants, both legal and illegal, are flooding into the U.S. because jobs are so plentiful.

Indeed, most inner-city youths say they can find jobs. If they are jobless, it is usually because they leave work often, not because they cannot find it. The long-term poor appear to decline work for personal reasons, or because they expect better pay than the market offers them. Overwhelmingly, the cause of poverty is not a shortage of jobs, or low wages, but that the poor do not work consistently.

Why is this so? Studies suggest that the long-term poor believe in the work ethic. Yet they have a hard time practicing it. This is surprising to most people, for the essence of middle-class behavior is consistency: One does what one intends to do. Mainstream citizens either obey social norms or flout them deliberately. Most analysts assume that the poor must be consistent too. Most poor adults, however, are less self-directed. Often raised in unstable homes with little discipline and few rewards for diligence, they need structure as well as help if they are to fulfill the work obligation they believe in.

This is the problem that workfare addresses and voluntary welfare reforms do not. Workfare does not reform the economy, or raise the skill or educational level of the poor. Rather, it closes the gap between the intentions of the poor and their behavior, and for this most recipients are actually grateful. Because the long-term poor already believe in the work ethic, workfare rests on moral authority more than force.

Over the last several decades, America has had considerable experience with voluntary welfare reform programs of the sort workfare opponents favor. For the most part they have failed. Voluntary programs are basically of two types: training programs and work incentives that subsidize the income of workers. Numerous evaluations of voluntary training programs by the Brookings Institution, the Urban Institute, the Congressional Budget Office, various universities, and other researchers have painted a remarkably consistent picture of their effects. Job training programs do improve the earnings of their clients but usually by less than $1,000 a year, not enough to lift clients out of poverty or off welfare. Even these small income gains often decay with time. More important, no voluntary training program has significantly increased overall work levels.

Badly designed training programs may even reduce work effort if they raise unrealistic expectations. Clients who have done poorly in school should not be encouraged to turn down available low-paid employment in order to pursue education or training plans from which they are unlikely to benefit. The “job skills” welfare clients need most are initiative, responsibility, and reliability—that is, the ability to get to work on time and take orders.

Even when voluntary training programs are well run, relatively few welfare clients bother to participate in them. In 1987, for example, New York City sent letters to 100,000 AFDC clients urging them to volunteer for training and employment programs; only 3 percent came to orientation, and only I percent participated. Even aggressive, well-run voluntary programs that offer extensive training and help with child care and transportation (such as Massachusetts’s Employment and Training Choices program) find they cannot reach beyond the most employable third of the caseload. Most of the participants who gain jobs would probably have done so without the programs. Voluntary training programs simply do not motivate most of the passive poor to do anything to help themselves.

The record of work incentives is even bleaker than that of voluntary programs. Work incentives seek to make work more attractive by allowing welfare clients to keep more of their outside earnings if they work, rather than deducting the whole amount from their grants. This reduces the welfare “tax” on their productive efforts. But studies by economists Frank Levy, Robert Moffitt, and others show that while stronger incentives may raise work effort among nonworking welfare clients, they reduce it among the working poor by allowing them to make the same money while working fewer hours. The net result, according to most studies, has been a decline in work effort for the low-income population as a whole, or at best little change.

During the 1970s, thanks to changes in federal law that allowed AFDC recipients to keep more of their earnings, the effective tax on those earnings fell substantially. Yet the percentage of welfare mothers reporting employment in a given month remained very low. The Reagan reforms in 1981 practically eliminated work incentives, raising the net tax on earnings of most welfare clients to 70 percent or more. But work effort by the welfare dependent did not decrease.

It should not be surprising that the long-term poor do not respond to economic incentives, for if they did, long-term poverty could hardly occur among the employable. Successful antipoverty programs must produce economic rationality in clients rather than assume it. That is why workfare succeeds where voluntary approaches fail.

Even by the standards traditionally used to assess job-training programs, workfare is more effective, though not markedly so. According to research by the Manpower Demonstration and Research Corporation (MDRC), workfare raises client income as much or more than voluntary programs, though still not by enough, in itself, to get clients out of poverty. It is also more successful than voluntary job training at getting clients into regular paid employment. In seven recent workfare programs studied by MDRC, the proportion of clients employed rose by as much as 37 percent (though from a low base). Workfare also reduced the welfare rolls in some cases, but to date this effect is not as pronounced as the increase in earnings and employment. (These estimates are for all those assigned to workfare programs, whether or not they participated; estimates confined to the actual participants would be higher.)

These traditional indicators, however, do not show the success of workfare at achieving its principal goal: inspiring more work effort from its clients. Workfare causes many more welfare clients to do something constructive to help themselves, whether working, looking for work, or entering training. It does this because the mandatory aspect of the program dramatically raises participation rates.

In the programs studied closely by the Manpower Demonstration and Research Corporation, about half the registrants participated in some work-oriented activity, including job search, public employment, training, or actual work, within six to nine months of entering the program. That is much higher than the participation rate for voluntary programs. In the best of these new workfare programs, in San Diego, 45 percent of clients participated actively within six months, compared with only 5 percent for comparable clients not in the program. Also, 25 percent of clients reported finding work during this time, compared with 13 percent for those not subject to the program. In the part of the city where the program tried to saturate the employable caseload with work activities, participation rose as high as three-quarters over a year.

Participation in turn leads to employment. My own statistical studies show that the higher the proportion of recipients that are involved in workfare, the higher the proportion that enter jobs and stay there. This is true even controlling for such factors as the employability of clients and economic conditions.

It is not necessary that the program put everyone immediately into a private-sector job search or unpaid work; education and training also produce job entries, provided they do not become a substitute for employment. But it is essential that clients be required to participate in something. Almost any workfare activity requires clients to get used to getting to work every day and on time. After this major adjustment, employment tends to follow.

Workfare, though mandatory, is much less punitive than advocates say. Because most recipients want to work, they usually accept the authority of the program. Though few clients will volunteer for job training, workfare programs find they rarely have to use sanctions once clients are told they must participate. The rate at which clients have grants reduced for non-cooperation is low even in the most demanding programs. In San Diego, 7 percent of mandatory job search clients and 22 percent of clients in mandatory work had to be sanctioned. The typical workfare program, however, still sanctions only about 5 percent of its participants.

According to Manpower Demonstration and Research Corporation surveys, over 90 percent of clients in mandatory government employment liked their jobs, and at least two-thirds looked forward to coming to work. Although the jobs were generally rudimentary, a majority of workers felt they had learned something. Large majorities usually said the work requirement was fair, and that they were satisfied to be earning their welfare instead of getting it for nothing.

On the federal level, welfare has been moving slowly in the direction of workfare since 1967, when the Work Incentive (WIN) program was enacted. WIN offered training and job-seeking skills and support. It was mandatory in the limited sense that most adult male recipients and some women were required to register with a WIN office. Then, if asked, they had to participate in WIN activities on pain of reductions in their welfare grants.

WIN had little impact, however, mainly because few registrants were called to participate actively, and those who were usually went into training programs instead of actual employment. In 1971, Congress reformed WIN to raise participation and work levels. In 1981, at the urging of the Reagan Administration, Congress allowed states to replace WIN with more demanding programs of their own. States were allowed for the first time to place AFDC clients in mandatory unpaid work.

Most states seized the chance, and new programs sprouted all over the country. Their success (as demonstrated by MDRC), along with the political popularity of work requirements, made workfare the theme of welfare reform in Washington. The 1988 Family Support Act (FSA) mandated that states sharply expand workfare programs. States are still allowed considerable latitude in the programs’ design.

But New York, with the largest urban caseload, has resisted workfare every step of the way. Part of the explanation is that welfare administrators, especially in the city, have been worried about other things. In the Sixties and early Seventies the city’s Human Resources Administration (HRA) was one of the most lax welfare departments in the country. By 1973, 27 percent of welfare claims in the city were being paid in error, compared with a national rate of 17 percent. Driven by threats of cuts in federal funding, HRA administrators put enormous effort into reducing the error rate, until by the end of the 1970s it was 10 percent or less. Under Mayor Koch, Blanche Bernstein and later administrators further strengthened the agency’s procedures. Today it is known for tough management.

The enormous effort to get and keep the agency under control, however, left little energy or political capital to spend on reform. Thus, although a succession of HRA administrators under the Koch administration proposed work-oriented reforms, they were unable to overcome substantial political and interest-group opposition in the city and in Albany.

More than in any other city, the nonprofit sector in New York is thick with foundations and service organizations that see it as their business to influence social policy. Most of these advocacy groups oppose workfare and most other restrictions on welfare. They have considerable political clout: easy access to administrators in both HRA and the state-level Department of Social Services, which has been particularly receptive to their views; close alliances with legislators in Albany, where the State Assembly has been the main roadblock to workfare; and considerable influence in the courts.

As a result of suits brought by advocacy groups, judges have imposed such detailed court orders on HRA that the Judges or the advocates themselves have in some instances virtually taken over policymaking. The ACLU, for instance, is now legally entitled to participate in city policymaking on foster care. Under pressure from lawsuits, the city not only agreed to house all homeless but to do so more generously, and expensively, than in any other city. The blizzard of lawsuits and special-interest lobbying smothers innovation. Reform requires leadership, but HRA is always playing defense. Reform requires the freedom to change; HRA constantly has new rules imposed upon it from outside.

HRA has never been able to make employment part of its routines. WIN required that employable welfare recipients sign up for possible job placement before they could get aid. I found in my studies, however, that New York’s welfare centers treated this as a formality. Clerks told applicants for AFDC that they had to “put in an appearance” at WIN, not that they were expected to go to work. Disgusted WIN staff dismissed the routine as no more meaningful than “getting your driver’s license renewed.”

This attitude destroys workfare. In effective work programs staff are supportive of clients, but they also enforce high expectations, because they believe the recipients can achieve self-reliance.

In New York I found two WIN offices, one in the city and one upstate, where staff rallied recipients with a combination of authority and enthusiasm. They levied obligations without flinching, but they also believed that clients had prospects and told them so. They treated recipients as morally responsible, thus shifting some of the obligation of overcoming dependency from the agency to the recipient. The aim, one social service worker told me, was that the dependent should “pay taxes and bitch about the government like everybody else.” Recipients hustled out of these offices to seek jobs as if shot from a gun.

Unfortunately, these offices were exceptions. The fault does not lie primarily with HRA. Starting in 1983, HRA initiated or proposed a number of work programs, all designed to break into the employable caseload and get more people to participate. One of these would have placed more recipients in unpaid work in exchange for their welfare grants, another in training or education. A third would have paid recipients higher grants for working. All would have required recipients to participate on pain of losing some portion of their grant.

In Albany, however, HRA’s workfare proposals were blocked and the agency never could institute workfare except on a very small scale. Under the Dinkins administration, workfare has largely been abandoned. The Cuomo administration has been unwilling to lead the fight for mandatory programs, preferring to launch a number of voluntary work programs.

In Albany the workfare debate intensified late in the 1980s as the U.S. Congress considered and then passed the Family Support Act of 1988, which came down firmly on the side of requiring welfare recipients to participate in workfare. By 1991 state programs must show that 7 percent of the employable welfare recipients (including mothers with children over three) are engaged in workfare programs on a monthly basis. By 1995, that figure must rise to 20 percent. States that fail to comply will lose some of their federal funding. These may sound like low participation rates. But because states must achieve them each month, in practice FSA will require states to enroll a majority of employable clients in workfare over the course of a year, or about the participation level achieved in the best recent programs. The act does not explicitly require that programs be mandatory, but no voluntary program has achieved participation this high.

Despite these new requirements, New York’s Assembly leaders continued to insist that the new programs be voluntary, stressing work incentives, training, and education virtually to the exclusion of actual work, and limiting workfare’s use of public jobs (in which clients receive no pay other than welfare) or sanctions.

The Senate, controlled by Republicans, wanted a program in which obligation was the keynote. Their plan would have required that recipients undertake a strenuous job search before qualifying for training. It was unafraid of unpaid work and tougher about sanctions. HRA generally backed the Senate approach, though its priority—correct in my view—was more to raise participation in a range of workfare activities than to insist that everyone work right away.

The most contentious question was higher education. The two houses argued over how long recipients could stay in college without actually going to work. The Assembly would have allowed a recipient to complete a four-year degree program before having to work, effectively turning AFDC into a scholarship program available only to the welfare poor. The Senate would have limited further study to two years.

In essence, the Assembly wanted a cosmetic version of workfare that might have generated enough success stories to give an impression that recipients were going to work, but without disturbing the bulk of the caseload. The Senate proposal, on the other hand, would have radically changed the nature of AFDC, firmly linking welfare and employment.

The quality of the debate was a disgrace to New York politics. In Albany, unlike Washington, neither the considerable body of research on workfare nor the national experience with workfare programs was seriously considered. On each of the specific disputes in workfare—unpaid work, job search versus immediate training, and sanctions—the hard evidence favored workfare.

Some of the advocacy groups were reasonable enough to negotiate with HRA over the shape of workfare. They would have accepted a greater degree of obligation in exchange for some new funding for education and training. The most influential groups and individuals, however, rejected workfare as a disingenuous scheme to evict clients from the welfare rolls.

The Advocates Workgroup on Welfare Reform, uniting 14 welfare lobbies, distributed a booklet in Albany advocating a very permissive version of workfare as well as much higher welfare benefits. The report deprecated any “harsh requirements or sanctions” as an affront to “equity.” One of these groups, Statewide Youth Advocacy, circulated a study claiming to show that HRA’s intentions in workfare were punitive. It accused the agency of “overuse” of unpaid work assignments simply to push people off welfare, rather than finding real jobs for recipients. It also accused HRA of “massive abuse of the sanctioning process.”

This document made much of the fact that unpaid employment and sanctioning are more common under HRA’s work policies than in other states, but it offered no evidence that this causes unusual hardship. Most likely, HRA more often uses sanctions and requires unpaid work because the New York caseload is unusually entrenched and has responded exceptionally poorly to voluntary work and training opportunities. Welfare practices in every state are now subject to uniform federal rules that protect clients. Past New York work programs have been decidedly undemanding, so it is implausible that HRA’s routines are punitive now.

The criticisms, nevertheless, were widely believed in the Assembly, especially on the key Committee on Social Services, in charge of the workfare legislation, which resolutely resisted the arguments of HRA managers. At one point, HRA sought the support of Charles Rangel, the influential Harlem congressman. Rangel voted for the Family Support Act but opposed HRA’s proposals after receiving advice from a number of welfare advocates, including James Dumpson, a former New York City welfare commissioner, and David Jones, president of the Community Service Society.

For their part, the Republicans who controlled the Senate failed to make their case effectively. Their version of workfare was more hard-line than experience or research suggests is necessary. The emphasis was almost exclusively on immediate )ob placement, but some recipients do need some training to be employable, especially in New York, where long-term dependency is so common and the culture of poverty so powerful. The better strategy is to expand involvement in a range of work-oriented activities, after which greater employment will follow. One must be severe about enforcing participation but lenient at first about the nature of participation. This was the style of the most exemplary work program to date, in San Diego. It was also HRA’s position. The Senate’s rather rigid emphasis on job placement was also impolitic, as it fed suspicions that workfare advocates were more interested in pushing people off the rolls than in helping them.

The debate dragged on long enough to cost the state $73.5 million in new workfare funding that was available under the Family Support Act during 1990. The combatants came to terms only in June 1990, just before the federal law became binding in October.

New York’s new program, called JOBS, is formally mandatory, but its teeth are dull. It limits the ability of welfare departments to impose unpaid work on clients, and limits the use of sanctions as well. It emphasizes assessing clients for remedial training before sending them on a job search. It does limit college to two years. Counties retain considerable discretion within these rules.

It is very doubtful that JOBS will satisfy the new federal requirement of 20 percent participation on a monthly basis by 1995. In theory, each county is supposed to meet the norms that apply to the state as a whole. That will be especially difficult for New York City. To comply, HRA would have to place many more recipients in job search and unpaid work than Albany has allowed it to up to now. The outlook is for further failure and conflict, with pressure building either to set a lower standard for the city, which would require higher participation upstate to compensate, or to get a reprieve from Washington.

New York’s leaders seem to feel morally compelled to do much more for the needy than other cities do, yet morally barred from demanding that recipients do anything to help themselves.

This ethos of entitlement stands between the poor and the only way out of the poverty culture. For their good and ours New York’s poor need to get to work; New York’s leaders need to stop discouraging them and help them on their way.


City Journal is a publication of the Manhattan Institute for Policy Research (MI), a leading free-market think tank. Are you interested in supporting the magazine? As a 501(c)(3) nonprofit, donations in support of MI and City Journal are fully tax-deductible as provided by law (EIN #13-2912529).

Further Reading

Up Next