Guidelines for State Welfare Reform
By Stuart M. Butler, Ph.D.
appreciate the opportunity to testify before members of the Illinois State Legislature on the im- portant topic of welfare reform. But the fact is you have far more to teach us in Washington than we can hope to teach you. Ever s ince the 1980s, it has been the states that have set the pace in wel- fare reform, and indeed almost all areas of domestic policy. That surge in state experimentation, made possible in part by greater flexibility to states afforded by the Reagan Administr a tion, has left the federal government vying to play catch-up. It has also meant that federal action and fed- eral laws often simply reinforce and extend the qtperiments that have taken place earlier at the state level. Unfortunately, as in the case of the 1988 Family Support Act, federal action often learns the wrong lessons from the states. And the federal government's efforts to encourage state innovation all too often simply frustrated it. In my comments today, I would like to share with you the finding s of more than ten years of state experimentation and scholarly work, and to suggest the lessons and guiding principles for ac- tion that flow from these findings. Adopting these principles, I should caution you, will not pro- duce overnight success, nor w i ll it lead to an "Illinois Miracle." But it will lead to steady progress in reducing the welfare rolls and improving the effectiveness of your welfare policies. To be sure, an "Iffinois Steady Progress" hardly has the ring of a "Massachusetts Miracle." Bu t unlike Gover- nor Dukakis's mythical improvements in the welfare system, Minois will be able to claim some- thing which is real. The experience to date points to six lessons about making welfare policy:
1) Our efforts to decrease material poverty have te nded to Increase behavioral poverty. Ile amount we as a nation have spent trying to alleviate poverty is truly staggering. Since the mid-1970s, total welfare spending at all levels of government, including the value of in-kind bene- fits that routinely ar e excluded from the official Census Bureau statistics on poverty, has averaged approximately 3.5 percent of the nation's gross national product. As my colleague Robert Rector has computed, governments at all levels in 1988 spent $184.2 billion on welfare p a yments, or $5,531 for each person in the U.S. officially classified as poor. Despite this, the official poverty rate has remained virtually the same, within a few percentage points, since the launching of the Great Society programs in 1965. During the 195 0 s and early 1960s, the poverty rate had been falling sharply. Our failure to achieve significant reductions in the poverty rate despite this level of spending is perplexing enough. But the impact of this heavy spending appears to have reinforced, or actua l ly caused, trends that have led to what has sometimes been referred to as a "culture of poverty," in which material poverty is made deeper and more persistent because of behavioral changes by the poor. One clear example of this is the reduction of work ef f ort related to welfare benefits. The past 25 years have seen a dramatic reduction in work effort among poorer Americans. There is convincing evidence that this directly related to welfare benefits. During the late 1960s and early 1970s, social scientists at the Office of Economic Opportunity (OEO) undertook a series of controlled experi-Stuart M. Butler, Ph. D. is Director of Domestic and Economic Policy Studies at The Heritage Foundation. Ibis is his testimony at a hearing held by the Illinois House Rep ublican Policy Committee in Chicago, September 24, 1991. ISSN 0272-1155. 0 1991 by 71be Heritage Foundation.
ments to examine the effect of welfare benefits on the poor. This study, known as the SeattleMen- ver Income Maintenance Experiment, or "SIME/DI ME," is the largest and most comprehensive such analysis ever undertaken in this country. The SIMF./DlME experiment found that every one dollar of extra welfare given to low income individuals reduced labor and earnings by 0.80 dollars. The impact was mos t pronounced among young unmarried males, with the number of hours worked declining by 43 percent. This pattern found in the SIMEVDIME experiment can be seen in the changing structure of the welfare popula- tion. In the mid 1950s, nearly one third of poor h ouseholds were headed by adults who worked full time throughout the year. With greater welfare available today, only 16.4 percent of poor fami- Hes are headed by an adult working full time. As Robert Rector has pointed out, "Thirty years ago the problem f a cing the working poor was low wages; today the problems is these adults don't work." Even more distressing has been that the expansion of the welfare state has coincided with the collapse of the family in low income households. In 1959, some 28 percent of poor families were headed by women'. Today that figure is a staggering 52 percent. And among black households, the illegitimacy rate in 1965 was bad enough at 25 percent, but today it is 65 percent-in other words, almost 2/3 of all black Americans bom tod a y are bom into a non-fwnily. While there are several factors behind these alarming trends, the effect of the incentives in the welfare system is the most pronounced. Ile welfare system today penalizes marriage and penal- izes work. To men struggling to ma k e ends meet to support a family on low earnings, the welfare system simply says, in effect, "If you want your kids to be better off, abandon them to the welfare system." And to the young woman who becomes pregnant, the welfare system provides just as it t e rse a message, "Don't work and don't marry. This sorry history of anti-poverty programs breeding poverty suggests two steps that should be taken by any state trying to reform its welfare system. The first is that welfare programs should be designed to ins i st on a two-way obligation. For 25 years the welfare system has been corrupted by the notion that while society is obliged to help the poor, the poor have no corresponding responsibility to use that assistance to improve their condition. This quarter cent u ry dominated by a one-way obligation is in stark contrast to the long tradition of public and private assistance being conditional on real efforts by the poor to improve their situation. This has meant that being on welfare can be an economic calculation b y beneficiar- ies, rather than always a last resort. That calculation has been as destructive for those sucked into welfare as it has been costly to the taxpayer. The notion of a two-way obligation, by contrast, leads to the idea that welfare should be li n ked to a requirement to work. While there should be some exceptions to that rule, such as in the case with women with pre-school children, the insistence that welfare requires work or a good-faith ef- fort to obtain work should be a basic element of all m a jor welfare programs, including housing as- sistance as well as Aid for Families with Independent Children and other benefit programs. The second implication is that welfare assistance should be explicitly pro-family rather than anti-family. This in turn s uggests several things a state should consider. One would be to copy Wisconsin's proposed experiment to reward welfare mothers who marry. It is not yet clear exactly how best to do that, but Illinois and other states should consciously try to make the wel f are system encourage rather than discourage marriage. Similarly, beefing up child support payments to discourage fathers from abandoning their families to the state also makes great sense. And re- quiring teen-age mothers to stay in the family home rather than setting up in their own apartment -which tends to mean long-term welfare dependency-should be a top priority for states. States can, incidentally, institute the residence requirements under current federal law.2
Furthermore, states should encour age efforts at the federal level to amend the federal tax policy to encourage work rather than welfare dependency. Ile recent expansion of the earned income tax credit (EITC) has helped. This refundable tax credit improves the earnings of low-skilled work e rs, and so makes work a little more attractive compared with welfare. Further expansions- of the EITC would give an additional incentive to work. A proposal for health care reform by The Heritage Foundation also would help encourage individuals to leave w e lfare and join the work force. Under this plan, the current tax exclusion for company plans would be replaced by a refundable tax credit for individual purchases of health insurance and out-of-pocket medical expenses. This means that a low-income worker i n a firm currently not providing health benefits would obtain government help to buy a basic health care plan. This would remove one of the major disincentives for welfare families to rejoin the work force, namely the loss of Medicaid benefits.
2) States n eed to think clearly about work requirements and training. While the idea of linking welfare benefits to work-or workfare-could now be considered the orthodox view among liberals as well as among most conservatives, there are many misunder- standings abou t what constitutes a good work/training requirement. As the Family Support Act in- dicates, many liberals have adopted the rhetoric of work requirements and yet flinch from actually placing those requirements into legislation. The Act seems on the surface t o be quite a strong in- ducement for welfare recipients to join the work force, but there are so many exemptions in the fine print that very few, if anyt individuals will be prodded into the work force. For the most part, the, "requirements"' simply are a d ditional benefits and training for those who wish to improve their skills rather than taking a job with their existing skins. To be sure, the Act win increase employ- ment, but the majority of the people getting jobs will be middle class service providers in the edu- cation and training industry. Our experience with education and training suggests a number of guidelines for meaningful work requirements. First, states should not focus on volunteers, as the new federal law allows them to do and as Massachuse t ts did under its Employment/Training program (ET). The assumption behind the em- phasis on volunteers is that the great majority of people on welfare would accept a job if only they had sufficient skills. Yet the overwhelming evidence suggests that motiva t ion is the problem, par- ticularly the lack of motivation because of the design of the welfare system itself. Thus focusing on volunteers and providing them with general programs to improve their skills generally does lit- de to encourage those resistent t o work to take a job. Indeed, those who might have otherwise left the welfare system for work often are induced to remain in the welfare system to receive free train- ing. Thus a real requirement to take a job is crucial to a successful work program. Seco n d, closely monitored job search programs seem to work somewhat better than other pro- grams. In other words, programs that take fairly elementary steps to increase the likelihood of em- ployment, such as encouraging welfare recipients to persist in follow i ng up job applications, or training individuals in interview skills, seem to result in greater success than elaborate skill-en- hancement programs. Third, New York University professor Lawrence Mead's exhaustive analysis of work programs suggests that not w ithstanding the slight edge for simple job search programs, intensive education and training, elementary job search programs, and other approaches have roughly the same level of effectiveness in getting welfare beneficiaries into employment. Moreover, the long-term earn- ings pattern of those enrolled in the intensive, and expensive, skills/enhancement programs is about the same as former welfare recipients who do not benefit from these programs before taking a job. Th us, from the point of view of impact for the dollar, the simpler the requirement the better.3
These observations suggest that states should focus on simple programs to increase the number of welfare recipients actually engaged in work experience or actual employment. In particular, the hours of participation required in a program should be increas ed to near full-time employment, perhaps thirty hours a week. Today, a participation period of ten hours is common. Moreover, the percentage of individuals on the welfare caseload should be substantially increased.
3) Understand the Incentive system by mea suring the real benefit levels. One of the biggest problems facing state and federal lawmakers in the debate over welfare is that the actual benefit levels available to individual households are rarely known by officials. When liberals argue for raising b a sic AFDC benefits, they tend to imply this is the only benefit re- ceived by the household. Yet there are about 75 federal means-tested programs, including often generous housing benefits, as well as food stamps, Medicaid, Head Start, energy assistance, a n d the Job Corps. In addition, states have a variety of programs, ranging from general assistance to medical, housing, and educational programs that supplement the federal assistance. Yet the Census Bureau ignores virtually all these in-kind programs in co m puting the income of those on welfare. In fact, of the $184 billion in total welfare spending, the Census counts only $27 billion as income for poor persons. Thus state welfare officials are, in most instances, flying blind. Not only do they rarely keep d e - tailed records of what the multiple programs used by individual families, but even ff they did, they would have no basis on-which to calculate the value of most federal in-kind benefits. In most states, officials have no idea what the benefits are. But t he recipients do! They know the effective value of fire housing, or free medical cam, and they make decisions whether to work or to remain on welfare on the basis of how these benefits compare with the cash income they would receive in employment. In many instances, families would be giving up as much as $20,000 in tax-free in- come in cash and in-kind benefits if the adults left welfare and took a job. It is not hard to imagine why many households decide that welfare is more attractive, even ff their basi c AFDC is much less than they could obtain by working. If states are to reform their benefit packages, they must know what they are. It is crucial for state officials to compute the real benefit values available to individuals on the welfare rolls. Only th e n can they make sensible policy decisions to assure that those who really need help get sufficient as- sistance, while those who are able to work are not discouraged from doing so by perverse incen- tives. And while this might seem a daunting task to many state officials, the basic work has been done. The U.S. Department of Health and Human Services has calculated the effective value of certain major benefits, as has the Census Bureau (although Census does not include these in its measurements of income). S imilarly, the White House made estimates of federal benefits for each state during the Reagan Administration. Thus there is a foundation on which to design a procedure for calculating the effective value for all benefits of a household. While a state is u n dertaking this necessary task of measurement, it is vital that it resists the cur- rent pressure to raise AFDC benefits. As I mentioned, the liberals tend to disaggregate programs. They point to the benefits of just one or two programs, such as AFDC, and t hen imply that this is all a household receives and that it is pitifully small. But the programs they identify actually are only the tip of the iceberg. It would be folly to raise AFDC without knowing the total package available to each family. To do so w ould cause even greater incentives for families to choose wel- fare over work and for fathers to abandon their families.4
4) Recognize the Importance of empowerment. There is increasing interest in Washington these days in a new term-empowerment. There are several threads to this idea. The first is that individuals tend to have a better idea of their own self- interests than any bureaucrat does. Thus giving peopI6 the means to run their own lives leads to better results than some bureaucrat vying to mic r omanage every decision of the poor. The second theme is that the very act of "helping" families by making every decision for them often turns them into passive individuals who become totally dependent on the government for their basic well being. This sen s e of utter dependency and inability to control events is perhaps the most perni- cious aspect of long-term welfare dependency. The third theme, related to this, is that insisting in- dividuals take responsibility for their own lives-be it with financial h e lp to do so-actually changes their behavior in a positive way. We have seen an example of these powerful effects in the movement for tenant management of public housing. Once tenants are permitted to run their own projects, and make decisions over ev- ery d ay aspects of their life in the project, their behavior changes and so does the neighborhood. In tenant-management prujects, rent collections rise, administrative costs decline, and the quality of life in the projects noticeably improves. Not only that, i n tenant-managed projects welfare depen- dency declines, teen-age pregnancy declines, and fathers return to their families. This happens be- cause in these projects there is both an opportunity to control one's life and an insistence by tenant leaders that each household lives up to some very clear responsibilities. Ne see a similar pattern in education in the rise of demands for choice in education. When once passive parents, who just accepted poor education for their children, are given the opportunity to make choices, they take the initiative. We saw this in the East Harlem School District, where re- sults improved dramatically once parents were able to make decisions over which school their chil- dren could attend. We see it even more dramatically today i n Milwaukee, where a voucher pro- gram for inner city children to attend private schools has led to demands by parents for a real shake up in the public school system. When individuals can make choices, they demand change and improvement. State officials n eed to examine the record of empowerment strategies. If they do so, they will quickly recognize that it is far better to provide individuals with direct assistance and the opportu- nity to make decisions than to fund service organizations to provide in-ki n d assistance to families. It is also cheaper. State officials should also appreciate that with Jack Kemp at the U.S. Department of Housing and Urban Development, and Lamar Alexander at Education, there are very strong proponents of empowerment at the fede r al level. Indeed, Jack Kemp now heads up a cabinet task force on em- powerment, and is willing and able to work with states to launch empowerment strategies at the state level. The great opportunity is available to any state wishing to try empowerment str ategies to break the cycle of dependency.5) Waivers are available to try now approaches. In keeping with the recognition that innovation invariably comes from the states, the federal government makes it possible, through exemption to federal rules, for s tates to try bold ideas. There are three ways in which this can be done. The first is through the provisions of the Family Support Act. This gives states the chance to de- sign programs to help individuals leave welfare by acquiring new skills. Unfortunat ely, the waiv- ers available through these provisions not only are limited in encouraging work programs, but in many cases the condition for receipt of federal funds is to avoid simple but effective strategies in
5favor of high cost skills-enhancement approaches. As I mentioned earlier, the latter approach is rarely as effective as simpler work requirements and is far more expensive. The second method of obtaining waivers is through Section 1115 of the Social Security Act. Al- though many state offici a ls are only dimly aware of this provision, it provides far greater discre- tion for state innovation than is available under the Family Support Act. More important, it per- mits very major changes in the AFDC program, allowing states to build in various w o rk require- ments and other obligations for the recipients of welfare. My colleague Robert Rector has suggested a number of waiver requests that states should con- sider under this section. One would be to introduce tough workfare requirements, including t he re- quirement that some AFDC recipients must work full time as a condition of benefits. Another would be to limit the availability of AFDC to no more than six years (basically the time required for a newborn to reach an age of full time schooling). The fact is if somebody is on AFDC for ten or fifteen years, the program has ceased to be temporary assistance and has become a way of life. Another experiment that could be conducted under the Social Security Act waivers would be some version of Wisconsin's L earnfare Program. In other words, receipt of benefits could be tied to the insistence that the children of a welfare household stay in school and avail themselves of an oppor- tunity to break the cycle of welfare dependency by obtaining a good basic educa t ion. In addition, some version of "wedfare" could be introduced, as Wisconsin is trying to do. This would provide incentives for broken families to become reunited, or for single mothers to marry without an im- mediate and significant reduction in benefit s . The strategy would be to encourage mothers on wel- fare to make the transition to becoming -a traditional family with a working father. Third, opportunity for waivers is still available under the approach first developed during the Reagan Administration . Under Reagan, a body known as the Inter-Agency Low-Income Opportu- nity Board was established to simplify the process of obtaining multiple waivers Erom different agencies. Essentially, a governor could come to the Board with a broad proposal for welfare re- form which necessitated several waivers from different departments. The Board, comprised of se- nior officials from every federal agency involved in anti-poverty programs, then would evaluate the proposal. If it agreed to it the Board would issue dire c tives to each agency to grant the neces- sary waivers. Thus it was a form of one-stop shopping for waivers. This Board helped to launch several major welfare reforms in the 1980s, including those in Wisconsin. Today that Board has been remodeled and now i s in the form of the Empowerment Task Force chaired by Jack Kemp. State officials considering welfare reforms should develop an overall strategy and present it to Kemp's task force. This would permit the approval of waivers in all the agencies concerned, a nd so would make very innovative reform possible.
6) There must be vigorous scientific evaluation and controlled experiments. As any scientist will tell you, experiments are only helpful under two conditions. The first is that there must be careful measur ement of the results. The second is that these results must be com- pared with a control group not subject to the changes made. Only in this way can the impact of re- forms actually be assessed. Unfortunately, it is quite common for states as well as the f ederal government to hail the success of a welfare innovations without accurately measuring its impact. This was the case in Governor Dukakis's once celebrated ET program. Dukakis claimed tremendous successes for the program. He was able to do so because h e permitted no systematic evaluation of ET. Indeed, attempts to evaluate it met fierce resistance from state officials, who for years denied scholars the most basic information. When eventually the program was systematically analyzed by the Pioneer Instit ute, a state-based research group in Boston, it was found to be a costly sham.6
It is imp ortant that other states do not repeat this "Imow-nothing" approach, either intentionally or unintentionally. Careful analysis and measurement, and comparisons w ith controlled groups of individuals not subject to a reform, is necessary for any state to feel confident that the apparent success of an innovation is real.The federal Food and Drug Administration insists that all new drugs be shown to be safe and ef- fective. I would suggest that state officials apply the same rule to changes in their welfare system. By experimenting on a small scale, an approach should be s hown to be "safe," meaning that it does not have damaging side effects such as increased dependency, high illegitimacy rates, and broken fmnilies-the side effects of today's welfare system. It should also be shown to be "effec- tive," meaning that a compa r ison with a control group shows that the reform actually does lead to an improvement and does so at a cost that is acceptable given the benefits. N states were to use this FDA rule in welfare reform, their efforts to experiment would lead to even better r e sults and more usable lessons for other states. The welfare system in America is a mess. It is an expensive mess. And it is a mess that is hurt- ing many families rather than helping them. Fundamental reform is needed. This requires a contin- uation and e x pansion of the healthy debate now taking place among welfare experts on both ends of the political spectrum. But these experts can only point to lessons that seem to emerge from pre- vious experience and they can only propose reforms. Moreover, experts wi t h exactly the same cre- dentials can differ markedly on their interpretation of previous experience and on their suggestions for reform. That is why state experimentation is so important. It is only by trying out proposals in good scientific fashion that g ood policy can be developed. That is why lawmakers in Washington should resist the temptation to try to solve the welfare problem by instituting grandiose new schemes that operate across the nation. To be sure, federal lawmakers need to address the basic i n- centives in the tax and benefit systems that discourage work and intact families, and foster depen- dency and broken families. But when it comes to the subtleties of the welfare system, and to designing detailed strategies that relate to the real condi t ions of families in particular neighbor- hoods, they should learn to take a back seat. They should provide information to states on what seems to work and what does not. Then they should remove the red tape of federal rules and give states the widest poss ible latitude to embark on radical reform.