Arthur Schiff is a research consultant in social policy and recently participated in an evaluation of several manpower training programs for the State of New York. The research for this article was funded by the Klingenstein Foundation.

To understand the current welfare mess, you have to understand people like Rose Johnson. Rose (her name has been changed) loves training programs. A mother of two boys, four and six years old, Rose has been on public assistance five out of the last seven years. This is the fourth time she has trained for a career as a word processor. Rose is no fool. She knows training programs can be a lot better than most jobs, and with the two kids to take care of, she has more than a day’s work at home. Besides, Rose has money problems. Unexpectedly, she had to go to North Carolina to visit her ailing mother. She was unable to pay the rent that month, so to keep her from being evicted, the welfare department paid it and is now recouping it from her welfare check; money is tight.

As a veteran of training programs, Rose is a smart consumer. Most training programs will pay stipends for transportation, lunch, child care, and even give a one-time grant for new clothes. The one Rose has just joined is a gold-plater. She has arranged child care for her boys with a neighbor whose kids she sometimes babysits in the evening. The welfare department will pay for the baby-sitting services; as part of her deal with Rose, the neighbor will split the child-care payments. Rose is bringing lunch from home and keeping the lunch money. She used the clothing grant for necessities and has managed to convince the department that she lives two fares away from the training site, although she actually walks one of the bus routes. When she finishes the program, she will go back to staying home with the kids.

Since the advent of the Great Society in the Sixties, America has experimented with a variety of programs to help the poor, from food stamps and expanded welfare benefits to job training. After almost 30 years both the taxpayers who fund these programs and the poor who use them are unhappy with the results. After more than 20 years of observing poverty programs up close, including nearly a decade as a player inside New York City’s welfare bureaucracy, I have slowly come to the conclusion that there is something fundamentally wrong with the way we think about helping the poor.

All of our diverse social service programs have one thing in common: They are based on a particular idea of what poor people are like: passive, helpless, dependent, and easily manipulated. This view of the poor is wrong, as is the usual portrait of the poor passively shuffling off to welfare centers to be signed up for the dole. Rather, most lower-income people are, like Rose Johnson, resourceful, practiced consumers of welfare and other social services. Like most consumers, the poor assess welfare "products" based on how they fit their own lives and goals. They perform "cost-benefit" analyses, exploring the benefits and requirements of assistance programs, deciding to use or reject various "products" according to their personal circumstances.

In the course of many years, first as a manager of one of these behemoth programs and then as a close observer of them, I have seen these consumer dynamics at work. Some people will not fill out a food-stamp application to get $200 per month in benefits; others will go to legal aid to fight for 10 dollars’ worth. Some people in dire straits will beg or borrow to avoid welfare; others, with a regular income, will hide it to continue getting Medicaid. I remember one older woman applying for food stamps. The interview was going well when the worker asked: "Do you have any money in a bank account?" "Goodbye," said the woman, getting up and stalking out.

For those truly interested in the welfare of the poor, this is good news, because it suggests that the poor are inclined to run their own lives and are mostly capable of doing so. The bad news is that social service programs elaborately designed to make the poor behave in certain ways, don’t. They simply become part of a stream of family income, a package poor people put together to live on.

The packages range from the unlikely strategy of living only on public assistance and food stamps, to the far more common strategy of combining income from as many sources—including work—as possible. Poor people, of course, break the rules and fudge the answers they put down on the forms. In this, lower-income people mirror the rest of society, which completes loan applications, mortgage forms, and job résumés with the same kind of creativity.

This is an important point, often overlooked by policy analysts. Government welfare policy fails in part because it is, inevitably, based on information furnished to the welfare system by the clients. Because poor people are not dumb, the information they give to inquiring welfare bureaucracies is hardly reliable. To get welfare you must tell the worst story about your life, about how you have no other money, no support from family or friends, no money from a part-time job, no income from unreportable sources. In this picture, welfare recipients really do look like passive, virtually helpless dependents, even "patients," whose lives can be micromanaged by benevolent bureaucracies for their own and for society’s greater good.

A recent, and very rare, in-depth study of the income and expenditures of 25 welfare mothers published in the journal The American Prospect painted quite a different and far more realistic picture. The study, by Christopher Jencks and Kathryn Edin, found that nearly all the mothers supplemented their welfare checks with income from other sources, including work: "None reported all her extra income to the welfare department, and only two reported any of it." National studies of the expenditures of welfare families, which turn out to be far higher than their supposed income, confirm this finding. So does common sense: Few mothers could support their families on welfare alone.

As this study suggests, the welfare system’s beneficiaries are not automatons or pawns; they are playing their own game. Thus, one of the most important and longstanding issues in the welfare debate—how we can get clients off welfare and back to work—is not a question about the real world. In the real world many or most welfare mothers are already working, though usually part-time or off-the-books, and less productively than if they did not have to conceal their income. They are on welfare because they want more cash and other benefits than they can get by working alone.

The official title of the basic welfare program is Aid to Families with Dependent Children. But for years the dominant theme of the welfare debate has been the misbehavior of parents and how to correct it. Welfare policy became preoccupied with such ambitious and complicated tasks as distinguishing the deserving from the undeserving poor and designing a system that would support those who could not work, without tempting those who could, balancing a decent provision for the poor against the lure of dependency. The system has affected the way the poor behave, but it has not directed them in the way social service professionals hoped. But because real poor people do not fit into these neat categories, the result was always a system that infantilized, demeaned, and corrupted the people receiving benefits.

The welfare system is a horrid, dispiriting failure for the same reason that Eastern European economies are horrid, dispiriting failures: Central economic planning does not work. Managing an entire economy and the lives of the people in it is just too complex a task for a central bureaucracy.

Poor people operate in their own special economy. Like all economies this one has its own rules; in this case, rules largely made by the social service bureaucracy. Welfare, food stamps, Medicaid, job training, school lunch, for that matter the public schools themselves, are closed-ended systems, worlds unto themselves, gulags if you will. Central planning replaces choice, forced work replaces jobs, waiting rooms replace motion, forms replace substance. Most importantly, unlike the rest of the American economy, the planned poverty economy is not accountable to its consumers.

There is a way to tear the iron curtain between welfare public policy and reality. This does not require "getting the government out of welfare" or abandoning our collective responsibility to the poor. The government, after all, has a very large role to play in the regular economy. Rather, the way out of the welfare mess begins by recognizing that the poor are customers. The demand for welfare and other services is a market. We must let that market work.

Businesses plan too. Businesses, like social service bureaucracies, must evaluate their clients’ needs and offer goods and services to satisfy them. But unlike central planners, businessmen never forget, because the market reminds them, that their customers are free to use their products or to find alternatives. In the real economy, businessmen cannot develop marketing strategies with the theoretical beauty and completeness of social science. Instead, they use a less elegant and ambitious, but far more successful method: market research. Market research may be as simple as watching which items move out of the inventory and which do not, but it is always driven by the humble realization that businesses must take the needs, desires, and goals of consumers into account.

The trouble with welfare is not that policymakers do not try to find out whether or how welfare programs "work." They do. Congress, state legislatures, welfare administrators, and all the experts who design social service programs rely heavily on social science research to tell them about the behavior of the poor and to design programs to modify that behavior. The problem is that social science research does not gather information as accurately or efficiently as a market.

Social science experiments are conducted on human subjects, a poor choice for laboratory tests. The vagaries of the human will are a variable for which no regression-analysis can control. Rose Johnson, our student of training programs, took part in an experiment to see if training would make her independent of welfare. It will not, not because the training program was objectively good or bad, well or inadequately funded, but because her goals are very different from those the job-training experts assume she has.

In the "laboratory," Rose Johnson wants to work and only lacks the panoply of supportive services (child care, counseling, role playing, group meetings, aptitude testing, basic education, health education, etc.), financial supports, training, and job placement services to achieve economic independence. These premises are wrong. What Rose wants is to stay home with her young children. The training program meets her needs for diversion and investment in the future, but she is not, in her own mind, in the market for a job. Experiments test hypotheses. But Rose Johnson confounds such experiments because she has her own agenda that defies the hypothesis.

Rose Johnson is in the program, essentially, because she was paid to enter it. But by paying her to enter the program, we deprive ourselves of the answer to one crucial question: If Rose really wanted a job, is this the sort of training program she would have entered and completed? If we had offered Rose no extra benefits, but charged her $10, would she have signed up?

Here the difference between social science research and market research is obvious. Market researchers would never bribe Rose to buy a product, because they want to know if Rose wants the product itself. But under the current approach to social services, we leave no stone unturned, no carrot withheld or stick unused, no rule unwritten, in an attempt to redirect the lives of the poor. Inevitably, though inadvertently, the system encourages clients to play games, as Rose did.

All of the successive waves of welfare reform—from programs to pay people to train, to programs to force them to work—amount to little more than failed attempts to do the hardest thing in the world: transform people’s hearts and minds to make them behave "better." The goals of market research, on the other hand, are more modest: to observe how people behave and to find ways to accommodate that behavior. Small wonder it is easier to sell a car than turn a criminal into a lawabiding citizen. The Rose Johnson scenario suggests we ought to keep it simple.

That was the conclusion of a recent $12 million evaluation of four work-training programs. All but one failed to achieve its goals. According to Mathematica, Inc., a leading public-policy research firm, the one that did succeed did so by jettisoning all the services and extra cash and simply offering skill training tailored to jobs available in the area. The time between entry into the training program and starting the job was therefore shortened; the trainees quickly learned the chore in order to get to work and to get the paycheck. The government paid for the training, which the employers might well have done.

Much expert opinion casts aspersions on such programs: The client was trained for a particular job, but not given the attitudes and background to fit him for advancement, or for other jobs; the client could just as easily have been trained to pick crops; when the job goes, the client will be back on welfare. Perhaps. But if we designed training programs on the basis of what people want, they would probably look a lot more like the one that worked than the ones that did not. The people we are trying to reach would say that classrooms are places of failure and frustration and what they want is a job. We would not spend time and money on Rose Johnson, or she on the program.

What about creating "the whole person," equipped to handle an uncertain future? Is that not a worthy goal? Actually it is not a "goal" at all. It is a social science paradigm. It does not make any sense to regard something as a policy goal if we have no idea how to achieve it, especially when the poor are able to frustrate any strategy for reaching it that we might devise. Distinguishing between the good we can do and the good we can only imagine is the difference between a training program designed by market researchers and one designed by social scientists. It is the difference, not only between two sets of values, and two assumptions about the nature of poor people, but far more importantly, if you care about such things, the difference between a program that works and one that does not.

In this and other cities around the country, professional advocates for the poor are a powerful force in designing and lobbying for social service programs. They are dedicated and generous people who work in social service agencies, public and private, and in more explicitly activist groups, such as the Children’s Defense Fund and the Community Service Society, whose primary mission is to lobby for expanded social programs. To change the system, these people, well-intentioned as they are, must change as well.

Advocates for the poor are rarely elected or hired by the people on whose behalf they speak. They hardly ever consult their poor clients while negotiating on their behalf and never submit those deals to clients for ratification. These advocates loathe and despise the same programs they fight to expand and impose upon lower-income people: They denounce food stamps as demeaning, but denounce even more loudly their underutilization; they decry welfare as a "bureaucratic nightmare," but want more people to enroll. Theirs is a difficult position to maintain. I know, I have been there myself.

It would dawn on my colleagues and me only dimly, when we were cutting deals for clients, that maybe they did not want a school breakfast program compared to, say, longer library hours. When an organization I was with fought for expanded emergency assistance to Supplemental Security Income (SSI) beneficiaries, we never consulted a single potential beneficiary. When we won, we celebrated our victory, not theirs.

There is considerable evidence that the programs we fought for so vigorously are not those the clients wanted. Or at least they did not want them in the way they were marketed, or with the elaborate, often discouraging strings that came attached. Major programs, such as food stamps, SSI, school breakfast, even welfare itself, are significantly undersubscribed. Simple programs like food pantries, often run by local churches, are besieged, while more ambitious universal assistance programs must advertise to attract recipients. Advocates blame bureaucratic obstacles as the cause of underenrollment, but thousands manage to use the programs even as thousands of others ignore them.

There is a market for loans, a market for mortgages, a market for insurance, a market for farm subsidies, and a market for labor. There is a market for food stamps, a market for unemployment compensation, a market for income transfers, a market for Social Security, and a market for school lunch. These markets have more in common than they do differences, if one has the minimal decency to credit all people—rich and poor, young and old—with the ability to make decisions about their own lives. Ignoring the information that markets can provide is a way to guarantee that programs will fail.

We want consumers—the poor—to run their own lives and do it better. And we want the system to offer services that might actually help the poor to do this, and to shut down programs that do not work. To accomplish this we must do two things. First, we must summon the political and moral humility to acknowledge that we cannot direct the lives of the poor in any elaborately detailed way. The poor need extra income, and we should provide it. But we must strip income supports of the detailed rules and bureaucratic oversight that inevitably encourage the poor to behave in ways that will keep them poor.

Secondly, we must design social service programs that respond to the market—to the people who use them. The great advantage businesses have over bureaucracies is information. If they guess wrongly about a product, the market soon corrects them. Welfare programs and those who run them need the same sort of feedback. We must build into social service "products" tests that will show whether the market is buying them. And we must shut down programs that do not prove themselves in the market, programs that fail to attract customers.

To shift from a bureaucratic model of welfare and social services to a market and consumer-oriented model would require a major change in the way the social service system and the academic community evaluate programs. Newly proposed programs should be pre-tested using market research techniques, such as consumer surveys, test marketing, and focus groups. Once programs are under way they must be made to face the acid test: If they attract only a small portion of their potential "market" because the clients do not show up, the programs must be allowed to die. For some programs, such as job training, we might even establish small fees to test the depth of customers I commitment to the product.

Even better would be to build market accountability directly into the programs, so that failed programs would die automatically, without the bureaucracy having to intervene. Currently, many social services are provided by private contractors, including private charitable groups and churches. Right now most of the work is done on contract. Wherever possible we should shift to a system in which the contractors must "sell" their services to the clients. Under such a system for, say, job-training programs, clients would receive from the government a voucher worth a certain amount of money, which the clients, with a modest copayment, could use to pay "tuition" at the job-training program of their choice. Of course, there would have to be safeguards against abuse, but the biggest safeguard is the market itself, the consumer preferences of the poor: Popular, effective programs would collect lots of vouchers, others would not attract enough clients to stay in business. Under such a system, the market research is "live," and the penalty for failure is bankruptcy.

Such tactics work best for service rather than welfare programs, which primarily provide income. Income programs also need a radical reform.

Our current welfare system is an unfixable failure. The only way out is to drop the requirement that people stay poor and dependent in order to get assistance. This will mean a huge philosophical change in the way we approach welfare. In fact, it will mean abolishing the welfare program as we have known it.

We can design income-support programs that give poor people the money they need, without managing their lives in detail or encouraging dependency. We can design programs that actively encourage work, not by rules and regulations around which huge enforcement bureaucracies must grow, but by built-in incentives that leave the details up to the people getting the benefits.

The principal task is to eliminate welfare and replace it with a national income-support system. The most obvious way to do this is to greatly expand the existing earned-income tax credit. The earned-income tax credit rewards poor people for having earned income, not for lacking it. It boosts their paycheck with a government bonus at tax time. An expanded version would say to potential recipients: As long as you demonstrate the basic initiative and social skills necessary to get a job and keep it, we will make sure you have enough money to live on. It does not have to be a good job, it can be any job. Just do that much, just do your fair share according to your abilities, and we’ll help you out. Instead of penalizing people for productivity, the way the current means-tested welfare program does, this program would reward people for productivity. As income rises, the benefits of the expanded earned-income tax credit would be gradually taxed away, creating a program that would tailor benefits to need, without discouraging work.

Inevitably, the program would end up giving some money to some people who by some definition do not need it. So does AFDC, even with tens of thousands of bureaucrats guarding the till. The advantage of the earned-income tax credit is that it eliminates the need for bureaucratic oversight.

Any government program has to have some sort of system of accountability, some way of knowing whether the program is working. But if we replaced welfare with tax credits, or even more direct wage supplements, we would not need bureaucratic accountability: The clients would be accountable not to the welfare bureaucracies trying to manage their lives, but to the real world, the real economy, as represented by their employers. No caseworker would have to check up on whether they were following through with their training programs, whether they had the right clothes for a job interview, whether they were developing the life skills they need for future independence. We could safely assume that their employers were requiring employees to do those things insofar as they were appropriate, and doing so in a less demeaning way than the social service system. The real market, not the poverty market, would provide all the enforcement the program would need.

The earned-income tax credit program would work with, rather than against, the market in other ways as well. Hundreds of thousands of jobs will materialize from the hidden economy when off-the-books jobs can no longer be used to complement welfare, and on-the-books jobs trigger tax credits. Even in the midst of a recession, low-wage industries are facing a labor shortage, because the wages they can afford to pay are not enough to lure people into work. If those wages were supplemented by tax credits, those jobs would become much more attractive. Of course, to keep the faith with job seekers in a world without welfare, the government could have to ensure that jobs are available if the private sector falls to provide them.

Some poor people, however, cannot or should not work. Of course, support programs for the disabled would continue. Some other nonwelfare programs, such as unemployment insurance, should also continue, as should a streamlined food-stamp program, stripped of dysfunctional work and household-composition rules.

Most importantly, however, once we get rid of welfare we will need a new program to help parents of young children. So to supplement the earned-income tax credit, the U.S. should also begin to do what many European countries do: pay parents, all parents regardless of need, child allowances to help with the costs of child-rearing.

Coming up with the right formula for a child allowance is a bit tricky. The program must be universal, because we do not want to require parents to stay poor in order to qualify. Equity, however, suggests that the government should not be supporting the children of the rich or upper-middle class, so the benefits (probably distributed through the social security system, to avoid establishing another bureaucracy) would also be gradually taxed away as incomes rise.

These reforms shift responsibility from economic planners to individuals. Because the child allowances and tax credits are not based on need, no means-testing bureaucracy and telephone directory-sized books of regulations will be necessary to manage them, and they should not encourage dependency. The allowance and the tax credit will replace not just the welfare grant, but the welfare client as well. The incentive to stay poor will disappear with the welfare program.

It is not hard to envision the outrage invoked by any attempt to deal with lower-income people as if they were ordinary human beings. Millions of benevolent Americans are more comfortable with the image of poor people as victims, or sick people needing counseling and care, or subjects for research, or objects for all the pathos, compassion, and religious concern they can muster. Others, less compassionate, see the poor as a body of people separated from the larger population by sloth, indifference, immorality, and self-destructive behavior.

What is needed is a third view of low-income people, which acknowledges their keen instinct for survival and their well-developed sense of self-interest. If we wish poor people to join the mainstream economy, we must allow them to play by the same rules that govern the mainstream economy. If we want to design programs that work, we must take advantage of the feedback mechanisms that make the mainstream economy work.

Poor people are no better or worse than the rest of the population—only ill-served as consumers in a nation that otherwise leads the world in its attention to consumer demand. To really reduce poverty we need to free the poor within our borders.


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