The first essential point about welfare reform is that its success shouldn’t be measured by this year’s income distribution charts. For most proponents of reform, the goal was, as Bill Clinton put it, to “break the culture of poverty” in America’s ghettos—to take a culture characterized by welfare dependence, a high rate of births out-of-wedlock, high male unemployment, and crime and replace it with a new, more virtuous social dynamic, in which every family would be expected to have a breadwinner, and consequently young men and women would make better choices about schooling, marriage, and childbearing. The test of reform’s success, then, is whether in the long run the largely urban, largely minority, welfare-reliant “ghetto poor” culture is absorbed into the mainstream American culture—whether “underclass” neighborhoods improve, employment rises, the out-of-wedlock birth ratio declines, the streets become safer, and children do better in school.
Reform proponents didn’t expect this to happen overnight. They were even perhaps willing to tolerate some short-term sacrifice of income among would-be welfare recipients if that was the price of a long-term cultural transformation. And there are some encouraging signs (the topping-out of the black illegitimacy ratio, the decline in teen births, the increase in female labor force participation, even the gentrification of former ghetto neighborhoods such as Harlem) that the desired change is under way.
That the 1996 reform also appears to have helped boost incomes in the immediate short run for most of the poor (see Ron Haskins’s article elsewhere in these pages) is also excellent news. But it remains true that a money test of welfare reform is a test imposed on the legislation by its opponents. The notion that the indicator of success is the number of Americans whose measured income (counting available government benefits) has been raised above some multiple of the poverty line—a widespread view propounded, most effectively, by the Center for Budget and Policy Priorities—is excessively narrow in several senses. It ignores the short term-long term trade-offs of the sort just discussed. It assumes that individual income is the overwhelming determinant of quality of life. And, most important, it arguably ignores the long-term goal of liberal politics, which I’d contend isn’t “more equal” income or “more equal” consumption, but fully equal respect and dignity—social equality. This is not a material concept. Ensuring a decent standard of living is one way to help achieve it, but it’s not the only way, and not necessarily sufficient. Work is another, necessary prerequisite: Americans who leave welfare for work gain the respect our society reserves for workers, even if they gain not a cent of income.
Politically, of course, low-income workers are a much more appealing constituency than welfare recipients. That was another goal of welfare reform, at least among Democrats: to transform the antipoverty debate into a debate about the working poor. The left has now discovered the political advantages of this transformation—without, of course, embracing the cause of the transformation, namely the 1996 reform. Still, now that new political possibilities have been opened up, what should be done to build on the 1996 law?
The Work-Ethic State
The American “welfare state” is not really a welfare state—in the American sense of “welfare—at all. Most government benefit programs have been “work-tested” since their inception. Workers’ compensation goes to workers. Unemployment compensation is conditioned on prior participation in the labor force. Social Security pensions go only to citizens who’ve worked a sufficient number of “quarters.” The earned income tax credit, which has ballooned into a large new entitlement, is by definition available only to earners.
It’s no accident that these big programs are work-tested; otherwise they wouldn’t have been popular enough to get passed by Congress. The two great exceptions to the general work test, Aid to Families with Dependent Children (“welfare”) and food stamps, are exceptions that “prove the rule,” as lawyers like to say. AFDC only came into being as an “entitlement” through a back-door process that involved a minimum of democratic approval by Congress and a maximum of intervention by the federal bureaucracy and the courts. Food stamps have an independent farm-state constituency. They’ve probably never been as unpopular as AFDC, but neither have they ever been as widely accepted and unstigmatized as Social Security pensions. Since the food stamp program persists, with only a partial work test, while AFDC’s successor, the Temporary Assistance for Needy Families (TANF) program, is haphazardly work-tested—and subject to a complex 5-year “time limit”—it makes sense to discuss these two exceptions first.
From TANF to a Guaranteed Job?
Given the booming economy, it’s been relatively easy for governors to push would-be aid recipients into the labor market and take credit for the resulting caseload decline. But policymakers will eventually face some harder choices. If there is a recession, recipients may start to pile up on welfare again. What are the states going to do about this, if the labor market isn’t providing ready employment for the unskilled? Even without a recession, the time limits built into the 1996 law may force a change of course. It’s true that the time limits are riddled with loopholes. (For example, states can declare that they are paying for additional years of aid with their own funds—a bookkeeping change.) But the publicity given the time limit has created an expectation, in the mind of the general public as well as of potential TANF recipients, that something is supposed to happen after five years.
There would seem to be three basic options. First, states can simply back off, using the existing loopholes to keep recipients on welfare even after five years. This would avoid any harshness, but also signal recipients that the time limit was not serious. And it would accept the presence of a permanent class of long-term dependents.
Second, states could send the opposite signal by strictly enforcing a time limit. The result might well be visible hardship—an increase in homelessness among families, for example—that would discredit welfare reform and the state governments in question.
The third, obvious alternative is to offer a community service or “workfare” job to everyone who hits the limit. This was, roughly, the approach taken in President Clinton’s original welfare reform proposal. It’s also the approach taken in Wisconsin—except that in Wisconsin the time limit is not five years, but effectively zero years. Wisconsin’s W-2 program offers community service jobs (rather than cash welfare) immediately to those who can’t find private-sector work. And Wisconsin’s program has been stunningly successful. Caseloads have fallen more than 90 percent since the late 1980s, and about 84 percent since 1993. Fewer recipients have wound up in workfare jobs than expected, allaying fears that workfare would turn into a large-scale “make-work” enterprise.
One of the great disappointments of the 1996 reform, so far, is that few jurisdictions have emulated the workfare-based Wisconsin approach. The time limit may change that. The mere threat of time limits seems to have altered the political battlefield, so that soon it may be the left that lobbies for mandatory workfare jobs, as a humane alternative to time-limited cutoffs. Unless the right moves the goal posts entirely and abandons workfare, agreement should be possible.
What do you call a system that doesn’t guarantee anyone a welfare check, but does offer a “community service” job? You could call it a guaranteed jobs program. Perhaps liberals should tactfully avoid saying that phrase out loud, but the underlying issue it raises remains: should a workfare job be an entitlement?
It depends on what the word means. “Entitlement” is, after all, a term of art that connotes a bundle of logically distinct program features. First, entitlements are available on demand to those who qualify for them. Second, they are funded automatically (at the federal level) as demand rises, bypassing the normal congressional appropriations process. And, third, they thus provide an automatic source of countercyclical spending that is directed at states with the greatest need. For lawyers, though, “entitlement” is also a magic word that creates a constitutional “property” interest, triggering a series of due process rights (and possibly equal protection rights).
In all except the last sense, I’d argue a workfare job is better off being an entitlement than not being an entitlement, although a period of state experimentation may be required before that conclusion is sensibly arrived at. But the last sense of “entitlement” is the most important to many advocacy groups, since an “entitlement” is a program easily tied up in litigation. The nub here may not be so much the procedures (hearings, cross-examination, and the like) required before an entitlement can be taken away from an individual, but the unsettled equal protection and uniformity issues. One county makes recipients clean the streets, while another lets them work indoors for fewer hours. One city provides two formal hearings while another provides only one informal session. One fires TANF recipients from their workfare job after a single fistfight or drug offense, another gives a second or third chance.
A federal court system that can find impermissible inequality in the disparate treatment of hanging chad can probably find impermissible inequality in this sort of variation, if what is at stake is the “property” right of an entitlement. For this reason—because it helps prevent legal aid lawyers and government employee unions (which tend to oppose workfare) from tying state officials in knots—the congressional denial of entitlement status to TANF was probably well-advised. I’d argue workfare jobs, even if guaranteed in practice, should likewise be freed from the nightmare of fairness that lawyers can create around “entitlements.”
Periodic attempts at work-testing food stamps make that program sound, in theory, less dole-like than TANF. In practice, the current food stamp work provisions go only part-way toward incorporating this benefit into a “work-ethic state.” For able-bodied adults without dependents between ages 15 and 50—the so-called ABAWDs—no-strings food stamps have, since 1996, been theoretically limited to three months every three years. But recipients are not required to work—they can participate in approved “training” instead. Even this requirement assumes training “slots” are made available to all recipients by the states. And states can exempt large groups of recipients, including those in areas with 10 percent unemployment.
For mothers with dependent children, the crucial loophole exempts those with a child less than 6 years old from having to work—effectively protecting almost 60 percent of TANF families. No wonder that, during the 1996 reform debate, food stamps were widely considered a reassuring safety-net for recipients who failed to comply with work requirements.
Because the food stamp program largely survived the 1996 reform, many liberals have seized on it as a platform for getting money to the poor—this time repackaged as a “work support” for those making the transition from welfare to work. President Clinton more or less went along with this campaign, even offering bonuses to states that did the most to “encourage food stamp applications.” But the campaign seems badly misguided to me. Food stamps don’t fit into a “work-ethic state” because they remain a form of welfare—they’re a cash-like benefit available to most single mothers whether or not they work. If many of the poor are reluctant to sign up for food stamps because they retain the stigma of welfare, it’s hard to say they are wrong. A low “take-up” rate is a good sign, indicating the presence of a work ethic, not a bad sign.
Attempting to sanitize a non-work-tested entitlement like food stamps sends the opposite signal from the one sent by the 1996 reform. If enough such welfare-like “work supports” become available, it will be possible to live (albeit uncomfortably) on the supports without the work. This isn’t a flip suggestion; in many Southern states the food stamp benefit already dwarfs the TANF benefit. And one lesson of welfare reform, even in high-cost cities like New York (where some welfare families routinely take the “sanction” of losing part of their check), seems to be that people are able to get by on far less formal income than previously thought. Of course, that amounts to the worst of both worlds—people living miserably in a state of dependence.
If incomes of workers need “supporting,” better to do it through a program like the earned income tax credit that is restricted to workers, or by an increase in the minimum wage. Nor would it be hard to create a separate food-stamp-like program that was actually, in practice, available only to those who join the labor force.
Earned Income Tax Credit
The message of welfare reform, then, is arguably this: “Programs to help poor workers should be restricted to workers.” The popular EITC stands as an ongoing rebuke to those social theorists who’ve adopted the familiar, competing slogan that “programs for the poor make poor programs.” The EITC is a program for the poor, and it works pretty well! The reason is that it vindicates the value Americans rightly place on work. Both political parties have attempted, over the years, to expand the EITC, until it costs more (around $30 billion a year) than TANF-style “welfare.”
The big question regarding the EITC—and other cash-like subsidies and supplements restricted to workers—is no longer how to preserve them against attack. The question is how high they can possibly be boosted before toxic side effects set in. Right now the EITC offers a bonus of $3,888 to a working parent with a two-child family who earns the minimum wage (about $10,700 a year). At some point—probably before the EITC becomes as large as the wage itself—the tax credit might have a strong disincentive effect on work as recipients decide that they can now make do with less labor. At some other point, the number of subsidies hinging on work might become so large that the idea that they are in any sense “earned” becomes politically insupportable. Also, increases in the EITC are becoming more expensive, since benefits must be phased out slowly as incomes rise, with the result that the tax credit becomes available at ever-higher income levels, extending into the more-populous middle class.
No Turning Back Now?
Likewise, work-tested subsidies other than the EITC—such as those for day care, transportation, and even health care—all, at some point, produce serious disincentive effects. What if that point comes too soon to achieve “social equality”? What if an increased minimum wage, plus the maximum feasible bundle of work-tested subsidies, is still too little, in and of itself, to permit a dignified life—especially in a society many of whose citizens are taking home heretofore unimaginable amounts? That wouldn’t necessarily be the end of the pursuit of social equality. I’ve argued elsewhere that society could still try to achieve that goal directly, by providing common social services (such as a uniform national health care system) to all or imposing common obligations (such as national service) on all—rather than indirectly by empowering the working poor to compete as individual consumers in the bazaar.
Even that might not work. It’s possible that the distribution of compensable human talents is so unequal that attempts to achieve any sort of fundamental equality of respect or dignity within a free economy are doomed. But I doubt that’s true, and I’m confident the vast majority of Americans doubt it’s true. It seems just as possible that maximum work-tested redistribution alone might do the trick. I’m not sure that even now most low-income American workers worry all that much about whatever social inequality flows from their financial status. Do our citizens really think that the average Wall Street stockbroker is better than the minimum-wage cashier down at the dry cleaner? Now give the cashier a $20,000 income, and health care, and decent schools, and safe, mixed neighborhoods . . . .
The American left, by promoting no-strings entitlements and denying the power of the work ethic, hit on the one plan guaranteed to disrupt this optimistic picture—by creating a class of isolated, stigmatized, dependent poor. The 1996 reforms took a major, perhaps decisive, step to dissolve this threat to equality. It’s no time to turn back.