u_topn picture
Atlantic Unbound Sidebar

Welfare: Where Do We Go From Here?

Opening Statements


Let me address Jack Beatty's questions one by one.

The real problems with welfare?

One of the main problems with welfare was that there was simply too much of it. There was too much welfare because we didn't -- and don't -- do enough to prevent the need for people to go on welfare in the first place, and because we didn't -- and don't -- do enough to help (and constructively push) people to get off once they are on. The prevention end is pretty complicated, but it includes the following: macroeconomic policy, the quality of public education, various other child- and youth-development efforts, serious enforcement of job-discrimination laws, economic development in poor areas, and approaches to strengthening the quality of life in those same areas.

As for welfare itself, it has never been aligned properly with the job market. Through policies such as the minimum wage, the earned-income tax credit, child-care assistance, and health coverage, people should be able to escape poverty by working. And while in general private-sector jobs are the preferable outcome, publicly funded jobs should also be created when necessary -- to assist with the transition from poverty and to take care of the increasing need for services such as child care.

Roundtable Overview

Host: Jack Beatty
Senior editor, The Atlantic Monthly

Sheldon Danziger
Danziger is a Professor of Social Work at the Univeristy of Michigan, where he is also the Director of the Program on Poverty, the Underclass, and Public Policy. He is the co-author of America Unequal (Harvard University Press, 1995) and the co-editor of Confronting Poverty (Harvard University Press, 1994).

Peter Edelman
A professor of law at Georgetown University Law Center, Edelman recently resigned his post as assistant secretary for planning and evaluation at the Department of Health and Human Services in protest over the welfare-reform bill signed into law last year by President Clinton. His article about that bill, "The Worst Thing Bill Clinton Has Done," appears in the March, 1997, issue of The Atlantic.

Robert Rector
A senior policy analyst of welfare and family issues for the Heritage Foundation, Rector's comprehensive welfare plan served as the basis for the "Real Welfare Reform Act of 1994."

Jason Turner
Turner is currently the director of capacity-building for Wisconsin Works, the state of Wisconsin's welfare-replacement program, much of which he designed. Prior to working for the state of Wisconsin, Turner was the director of family assistance at the U.S. Department of Health and Human Services under President George Bush.

The new system is going to make things worse. Instead of tackling the hard issues and making necessary investments, Congress and the President have thrown the issue at the states with incentives and requirements to push people off welfare arbitrarily.

Robert Rector responds:
"Peter Edelman talks of making 'necessary investments' in new government programs. This is an interesting phrase, especially since Lyndon Johnson in launching the War on Poverty called that program an 'investment which will return its cost to society many fold.' Our nation has since that time invested more than $6 trillion in fighting the war on poverty -- and virtually every social problem has gotten worse, not better."

See the rest of Rector's response.

The new law doesn't promote work effectively and doesn't protect children. The old system involved at least a framework that was right: a national definition of eligibility for assistance and a national assurance of a federal share of the cost in both good times and bad. If you applied for help and you met all the conditions, you had to receive assistance promptly, and if you were pushed around you were assured of a hearing to look into the issue. That framework was sound. It should have been fixed, not dynamited.

The worst single thing about this new law is that the states have no legal responsibility to help anyone. They can do real reform if they are willing to put up the money -- but they can also, as Jason deParle of The New York Times has written, give people a bus ticket out of town. When the time limits hit (and some of the states are adopting time limits even shorter than the federal five years), and if a recession hits, I am afraid we are going to see some real suffering. I would be somewhat less disturbed about that if children were not involved, but most of the sufferers are going to be children.

Does real welfare reform cost money?

In a word -- yes. In fact, it costs money not just to get people off welfare but also to help the increasing number of people who struggle to stay off welfare but can't make enough money to get out of poverty. The Wisconsin model conceptually goes in the right direction by saying that

Robert Rector responds:
"No matter how much spending is increased, liberals always come back complaining and demanding more. This just isn't the way to go. Welfare, by undermining the work ethic and marriage, insidiously creates its own clientele -- the more that is spent, the more people in apparent need of aid are generated."

See the rest of Rector's response.

our approach should be job-based, but it looks to me like it isn't really investing in getting people out of poverty. I'm very troubled that the community-service "jobs" it creates don't pay the minimum wage and are not defined legally as jobs, with the result that the earned-income tax credit doesn't attach to them. I'm also worried that Wisconsin's approach doesn't take sufficient account of people's individual circumstances. Some people are on welfare because of personal and family problems that aren't going to go away. Wisconsin looks good right now because they have a very hot economy. We'll see how things look down the road.

Chronic dependency?

Something like half the national welfare caseload on any given day has been on the rolls for more than five years. That's not half the people who ever go on welfare but rather half the caseload at any given time. Most people who go on the rolls get off more quickly, although many come back on again. But yes, there are too many who go on and stay on. That's really what this whole debate is about.

My position is this: invest in prevention and rearrange the carrots and sticks for those who do end up on welfare. What has been enacted is a gigantic cop-out that is going to hurt a lot of people, especially children. Many of the people who enacted this law believe in good faith that they have committed an act of tough love. I think they have made a terrible mistake. I'm not going to get into the remark by Congressman Shaw that Jack Beatty cites: if Shaw actually said what he is quoted as saying, I deplore it, and I would guess that he regrets having said it. He is one of the people who genuinely believes that this bill was the right thing to do. It suffices to say that I think he is wrong.

David Ellwood?

David Ellwood is not responsible for this new welfare law. The President and Congress are. David Ellwood came up with a constructive proposal to make the kind of public investment that would alter the incentives and get people into jobs in a positive way. This law perverts his proposal and is not even based on the research we have about policy outcomes. In fact there were a large number of new experiments going on and being evaluated in the states under the waivers that HHS had granted in recent years, but the President and Congress rushed to judgment anyway. A federal welfare reform enacted in 1988 was also still being evaluated. What we've now got instead is a law that takes the knowledge we have about chronicity of receipt of welfare and makes a gigantic leap of faith -- if we can call it that -- concerning a totally untried, underfinanced, and most importantly, undefined forced march of poor children out onto a high wire without a safety net. The passage of this new law tells us what we already knew: politicians make decisions that are not based on research and experience.

The logic of the block-grant idea?

Don't ask me. I believe in more flexibility and innovation in the way we spend federal money in many areas, and welfare itself could have used more flexibility along the lines of many of the welfare-waiver proposals that had been put into effect. But a federal role in some basic definitions of income support for poor people is appropriate along with federal matching funding based on how many people qualify for help at any given time. I just don't see the logic of the block-grant idea as applied to welfare.

Where do we go from here?

Not to Congress. They are not about to do anything major to fix things. All we can do is focus on two areas: the implementation of the law at the state level and the collection of the data necessary to document what happens.

Return to the top of Round One -- Opening Remarks.


In order to answer Jack Beatty's opening question and explain why the new welfare bill won't have the catastrophic effect

Peter Edelman responds:
"Despite Turner and Rector's rosy confidence, we liberals -- if it makes Turner happy to use the 'L' word -- actually know something from real-world experience about how hard it is for long-term welfare recipients to get and hold jobs, even when they indisputably want to go to work and keep on working."

See the rest of Edelman's response.

many liberals fear, I would like to start by looking at the central premise of Peter Edelman's cover story. The article -- which contains the pantheon of liberal assumptions, fears, and misconceptions about real reform all together in one place -- essentially argues that poverty is the cause of the urban maladies about which we are all concerned.

Edelman says that the reform bill will result in more crime, malnutrition, infant mortality, domestic abuse, and drug and alcohol abuse. Whew! To respond responsibly we must first acknowledge that low income is associated with adults who don't work, who are substance abusers, and who commit crimes, among other things. But is low income in and of itself the cause of these social maladies?

If the answer is yes, then one must stretch one's mind to accept that the following are all true as well:

  • Welfare-dependent families in California -- where families of four receive a higher combined income from food stamps and AFDC than do full-time minimum wage workers supporting families of the same size -- have fewer social problems because they have more money.

  • The welfare-dependent population in Los Angeles has fewer social problems than the same population in Chicago because AFDC benefits in California are almost twice as much as in Illinois.

  • Urban families in poor countries throughout the world have higher levels of social dysfunction than those in the United States.

  • Social problems were worse for the average family during the Great Depression than for welfare-dependent families now, since the majority of the former had lower average real-income levels.

Come on, get real! The truth is that in a rich country it is not the relative level of income that is a prime social determinant, but rather the source of income, by which I mean either self-generated

Sheldon Danziger responds:
"Mr. Turner and Mr. Rector write as if the labor-market problems of welfare recipients take place in a vacuum, and as if the primary problems welfare recipients face are due to the rules of the AFDC program. I could not disagree more."

See the rest of Danziger's response.

or donated. It is enforced idleness that is responsible for much (though not all) of the decay in the social and family fabric. The recent welfare-reform law, through its work requirement, begins to make real the promise of the earlier Family Support Act -- that there be a system of mutual obligation between the needy and society at large. It is a good law, and states such as Wisconsin that intend to use it to re-create a viable system of social obligation will see healthier families and communities as a result.

Does real welfare reform cost more money, and is this the lesson of Wisconsin? The answer is that real welfare reform almost always brings down overall costs because it reduces caseloads as families go to work. Wisconsin Governor Tommy Thompson frequently speaks about the additional investments he has made to help welfare-dependent families escape the system. But one must understand that he is referring to increased investment per family. Because the reforms in Wisconsin have resulted in the caseload's being cut in half since Thompson came to office in 1987 (and cut by almost a third over just the past twelve months), the resulting overall budget for AFDC, job training, and child care is now far lower than it was previously.

What about the question of the role of research in welfare policy-making? In l995 I finished leading the state effort to design a new work-based family-assistance system to replace AFDC. This became Wisconsin Works (W2). While we were able to glean some general notions from the available good research (for example, a work-first program orientation is more effective than an education- and training-based one), the truth is that much of the academic research being done today has no practical value to those designing programs. Most of the research relies on calculations from data sets that show the association between one factor and another (for example, poverty is associated by x factor with illegitimacy). This kind of research is completely useless because it cannot relate cause and effect (ergo Edelman's misconceptions). Research played only a tangential role in the development of Wisconsin Works.

Moreover, one of the most important tasks for welfare-to-work practitioners -- improving the effectiveness of the institutions and providers that actually deliver welfare programs, particularly in urban areas -- is not even a subject of interest for most academic researchers, with the notable exception of New York University professor Larry Mead. The service-delivery system for welfare programs introduces a severe limiting factor for designers of social programs that is not generally understood by the public or most academics. We have attacked the problem of improving institutional capability in Milwaukee by dividing the city into six manageable sub-units, by conditioning agency budgets upon meeting performance standards, and by permitting private vendors to compete with public-sector agencies.

Research that is based on studying incremental changes to the old AFDC system can only be so useful in today's world. If the times call for bold or even radical change, then we must move ahead forthrightly and responsibly, but perhaps without a full complement of experimental research to guide us. This decade's great systemic changes will provide us with plenty of ideas that will help shape the future.

I close these remarks with the following observation. Liberals as a group are predicting that the states, concerned about caseloads and freed from the

Peter Edelman responds:
"While I'm glad to hear that Wisconsin wants to raise the pay for community-service jobs, what's being proposed is still way less than the minimum wage, and I still want to know why Wisconsin doesn't attach the earned-income tax credit to those 'jobs.'"

See the rest of Edelman's response.

federal-entitlement obligation, will cut benefits and reduce the numbers of those deemed eligible for benefits. Two weeks ago, however, Governor Thompson released his proposed budget for the next cycle. In it he proposed an increase in the benefit amounts to be paid to those working in community-service jobs under W2, the future welfare-replacement program. That's a twenty-percent increase that will raise the monthly benefit for working in a full-time W2 community-service job from $555 per month to $678. This is the same governor who cut AFDC cash benefits upon entering office in 1987 and who allowed them to remain flat until now.

Why the change? (Note: the Governor's welfare-department staff did not suggest it.) The answer lies in the nature of the benefit. Under W2, benefits will be granted only in exchange for work, and that makes all the difference to a conservative governor -- and to the public.

Return to the top of Round One -- Opening Remarks.


Jack Beatty raises a number of questions in his opening statement that deserve attention -- more than I can address in this space. I'd like to limit my focus to the questions of why real welfare reform and budget cuts are incompatible and what will happen when a recession hits.

There was bipartisan agreement in the early 1990s that we should transform the Aid to Families with Dependent Children program; the idea was to move from a cash-based safety net to a work-based safety net by transforming the entitlement to a welfare check into an entitlement to the opportunity to work. The Personal Responsibility and Work Opportunity Act of 1996 (PRWOA), however, will not achieve this transformation. Indeed, its title is a misnomer, as it provides no work opportunities, or any assistance at all, for recipients who reach its time limits. The new law just ends welfare and absolves the government of all responsibility to care even for those who want to work but cannot convince an employer to hire them. The PRWOA cuts total welfare spending even though research suggests that additional funds are needed if most recipients are to get and keep jobs.

Many states will find it difficult, in Candidate Clinton's 1992 words, "to end welfare as we know it.... without punishing the poor" unless they spend substantial additional funds of their own. Welfare-to-work programs are expensive because of an economic mismatch between the kind of labor that can be supplied by welfare recipients and the kind of labor employers are demanding.

On the supply side, many welfare recipients have limited educational attainment, and job skills and are disproportionately located in poor, inner-city

Jason Turner responds:
"Danziger correctly points out that many employers require high-school diplomas, experience, and references. This is because they have been burned in the past by unreliable workers. After closely questioning many employers who require high-school diplomas, I have come to the conclusion that many or most are using the diploma as a proxy screening device for other qualities -- especially persistence and responsibility."

See the rest of Turner's response.

neighborhoods -- where there are few job opportunities and from which they have difficulty commuting to available jobs. On the demand side, employers during the past fifteen years have escalated their expectations of workers. They require high-school diplomas, experience, and references. About half to two-thirds of all newly available jobs for which employers do not require a college degree involve each of the following on a daily basis: the reading of paragraphs, the performance of basic arithmetic, dealing with customers, and the use of computers. Only about 5 to 10 percent of new jobs require none of these skills. Because of these increased skill demands, even many young male high-school graduates who do not have child-care responsibilities have been faring poorly in the labor market.

Given the characteristics of welfare recipients and changing employer demands, the most successful state welfare-to-work programs -- most notably, the GAIN

Robert Rector responds:
"Sheldon Danziger refers to the GAIN program in Riverside, California, to demonstrate that work programs can't really reduce dependence. I am sure that Jason Turner got the same chuckle out of this as I did. Conservatives have long recognized that Riverside was a flimsy, permissive program; it represents the minimum effect that can be expected from reform, not the maximum, as Mr. Danziger claims."

See the rest of Rector's response.

program in Riverside, California -- rarely obtain long-term employment rates of more than 50 percent. Many recipients have difficulty obtaining any job offer, even if they search extensively. Others obtain work but frequently lose jobs owing to poor preparation or performance, absenteeism caused by the demands of child-care, child-health or transportation problems, or the sporadic nature of the low-wage job market.

Michigan terminated its General Assistance (GA) program in 1991; the experiences of recipients who came up against this "time limit" suggest that ending cash assistance will not get most former recipients into jobs. A sample of GA recipients was interviewed two years after the termination: among those who were under the age of 40 and had not qualified for disability benefits, about two thirds had worked at some time during the two years, but only one half of those with a high-school or GED degree and one quarter of those without degrees were employed at the time of the survey.

Welfare recipients should have the personal responsibility to look for work, but if they diligently search for work without finding a job, they should not be cut off from all cash assistance. At a minimum they should be offered an opportunity to perform community service in return for continued cash assistance. Fear of destitution is a powerful incentive to survive. It does not, however, guarantee that an unskilled worker who actively seeks work will be able to support her family over the long run. The harsh realities of the labor market mean that restricting assistance for welfare mothers will increase economic hardship.

The likelihood that the new welfare law will cause harm will increase over time. The labor-market prospects for welfare recipients are not good today, even with unemployment below 6 percent. They will surely get worse during the next recession. And when unemployment rises the states will not be inclined to respond, because of their fixed federal block grants. Only the rare state will

Jason Turner responds:
"Danziger opens by saying that real welfare reform and budget cuts are incompatible. He concludes these same remarks by saying that what is really needed are wage supplements for low-income families and jobs of last resort for those who cannot find private employment. Wait a minute! Both of Danziger's solutions have already been enacted into law. Where's the beef?"

See the rest of Turner's response.

increase welfare spending during a recession, when tax revenues are declining and the budget has to be balanced. Indeed, the most likely prospect is that states will cut benefit levels and contract their time limits during the next recession.

The only way to reduce both welfare dependency and poverty is to institute a set of labor-market policies that would both supplement the wages of low-income families and provide low-wage jobs of last resort to those who seek, but cannot find, employment. This requires a Mutual Responsibility Act and a willingness to spend public funds for a work-oriented safety net.

Return to the top of Round One -- Opening Remarks.


The following are my responses to the questions posed by Jack Beatty in his introduction:

Problems in welfare?

Welfare costs a lot and generates dependence and illegitimacy, both of which are harmful to children's development.

Sheldon Danziger responds:
"Mr. Rector acts as if welfare receipt itself has a large causal role in child development. I would argue instead that poverty has negative effects on child development. Poverty is the causal factor."

See the rest of Danziger's response.

Altogether, the federal government and the states spent $404 billion last year providing means-tested welfare assistance. This sum included cash, food, housing, medical care, social services, and targeted education and training for poor and low-income Americans. It amounted to a record 5.3 percent of the GDP. Roughly half went to families with children.

The welfare-reform law will slow the future growth of this spending very slightly. It will significantly reduce dependence by requiring welfare recipients to work in exchange for benefits, which in turn dramatically reduces the attractiveness of welfare and increases the likelihood that recipients will leave welfare and stay off. The bill also provides new incentives to reward states that succeed in reducing illegitimacy without increasing the abortion rate.

Costs more?

The notion that welfare reform would increase costs has always been a blind article of faith among liberals which conservatives have found quaint. The goal of reform is not to shove welfare recipients into make-work public-sector jobs. Rather, requiring welfare recipients to perform community-service work in exchange for their welfare checks causes recipients to leave welfare in droves and vastly reduces the numbers applying for welfare in the future. Very few recipients actually remain in community service for long. This has already been demonstrated in Wisconsin, Oregon, Massachusetts, Utah, and other states.

Length of dependence?

The typical family now on AFDC will receive benefits for about thirteen years. Once they leave AFDC, they will generally continue to receive other kinds of welfare benefits.

David Ellwood?

David Ellwood had no role in the welfare-reform law. By contrast, the Wisconsin work programs implemented by one of our panelists, Jason Turner, had a very great bearing on the crafting of the federal law.

David Ellwood did vaguely popularize the idea of time limits for welfare. However, the importance of time limits to the federal reform has been greatly overstated. In fact, the time limits in the new law are merely symbolic. True, the law does create a five-year time limit, but this limit is deliberately applied only to federal funds -- not to state AFDC money. Why does this matter? A little arithmetic will help: about half of AFDC funding comes from state sources that are unaffected by the five-year time barrier, and historically about half the AFDC caseload at any given time has received aid for more than five years. States may thus avoid the time limit merely through an accounting change. They may give state funds to recipients who have been on welfare for more than five years and give only federal funds to those on welfare for less than five years. A little fiscal juggling and presto! -- the allegedly draconian time limit vanishes completely.

Both federal and state policy-makers understand that the "time limit" may be evaded in this manner. Although I do not expect states to ignore this symbolic limit completely, the federal law does provide vast flexibility to continue cash welfare beyond five years whenever a state feels this is needed.

Logic of block grants?

The old AFDC program was funded on an entitlement basis: the more people a state enrolled in AFDC, the more funds received from the federal government. Conversely, if the number of people on welfare was cut, federal funding was reduced. This "entitlement" funding structure created perverse incentives for state governments, penalizing states that reduced dependency and rewarding states that allowed their welfare caseload to grow. The welfare-reform legislation eliminates these perverse incentives. Under the new law, each state will be given a fixed amount of money that will be gradually increased. If a state reduces its AFDC caseload, its federal grants will not be cut; states may then use surplus funds to provide more intensive services for the poor.

If states implement serious work requirements like those in place in Wisconsin and Oregon, their caseloads will plummet (while AFDC funding remains fixed). This will create a substantial fiscal cushion by the time of the next recession. Moreover, Wisconsin shows that once real work requirements are in place, AFDC caseloads do not increase, even during a recession.

Where do we go from here?

The liberal defense of the welfare status quo is based on five myths: 1) welfare recipients want to work but no jobs are available; 2) a shortage of

Peter Edelman responds:
"I'm fascinated by the confidence of Jason Turner and Robert Rector that this massive kick in the pants is going to magically do the job. About all we know right now is that the welfare rolls have been going down coincident with unemployment's being low for a fairly sustained period of time and coincident with welfare offices in some places responding less favorably to applications than before. "

See the rest of Edelman's response.

daycare and transportation makes employment unlikely; 3) expanded education and training are the key to reducing dependence; 4) rapid reduction in welfare caseloads will yield severe economic deprivation; and 5) reforming welfare and requiring recipients to work costs more than the old system.

Each of these myths has already been decisively disproved by reforms in Wisconsin, Oregon, Oklahoma and other states. The bottom line in welfare reform is simple: If recipients are required to perform community-service work in exchange for their welfare checks, they will stay away in droves.

Among the states, Wisconsin and Oregon have the most rigorous work requirements, and caseloads there are vanishing so fast it is difficult to keep track. Wisconsin has already reduced AFDC dependence by 55 percent, and the welfare caseload continues to fall, by 2 percent per month in inner-city Milwaukee and 4 percent per month in the rest of the state. Other states must enforce serious work requirements as well.

Roundtable Overview

Introduction and opening questions by Jack Beatty

Round One -- Posted March 12, 1997

Round Two -- Posted March 25, 1997

Copyright © 1997 by The Atlantic Monthly Company. All rights reserved.
Cover Atlantic Unbound The Atlantic Monthly Post & Riposte Atlantic Store Search