A welfare mother in Central Harlem is not poor for the same reasons that a subsistence corn farmer in Mexico is poor. That’s just one of the many self-evident conclusions to emerge from a dangerously misguided antipoverty program begun by New York City Mayor Michael Bloomberg in 2007. Bloomberg’s initiative, Opportunity NYC–Family Rewards, bestows cash rewards on (for the most part) single parents and their children if they act responsibly—by attending school, for example, or by working. Interim results for New York’s effort, which is modeled on a program in Mexico that targeted agricultural peasants, are now in. Not all of them are as blindingly obvious, though, as the fact that multigenerational urban poverty in America is far different from Third World rural poverty. One detail in particular stands out: that the lives of America’s underclass are characterized by a degree of disorganization that is rarely grasped or acknowledged. The implications of this revelation will be difficult, if not impossible, for the welfare industry to accept.
The much-heralded initiative that inspired Bloomberg’s program was based on the theory that Mexican campesinos were so trapped in the daily struggle for survival that they couldn’t undertake behaviors that would help them escape poverty over the long term. Facing grinding economic pressures, parents pull their children out of school to help with the harvest; mothers don’t take their children to the doctor because they can’t wrest time away from work in the fields or at home. Oportunidades, as the Mexican initiative was called, tried to change the perceived zero-sum relationship between self-improvement and present income by paying Mexican farm families for taking actions in their long-term self-interest, such as school work and medical visits.
Bloomberg’s version of Oportunidades officially pretended that New York’s underclass faced similar tragic choices. The poor failed to “plan for the future” because they were “so focused on surviving,” the mayor explained when inaugurating the program. They were engaged in a “struggle” for the very basics of existence, he said. Opportunity NYC–Family Rewards would alleviate the immediate pressure for survival that allegedly prevented the inner-city poor from investing in their future by paying them for self-improving behavior, thus offsetting the purported opportunity cost of future-oriented actions.
Of course, it’s ludicrous to suppose that what keeps America’s inner-city residents poor across generations is a struggle for subsistence in an economy of limited opportunities. The main drivers of poverty in America are family breakdown (in 2004, single-parent households nationally were six times as likely to be poor as married families) and nonwork (only 5 percent of all families with one full-time worker were poor in New York City from 2005 to 2007, compared with 47 percent of families with no workers). The antisocial behaviors that contribute to multigenerational poverty also have nothing to do with suffocating economic pressures: very few inner-city students cut classes or drop out of school to help their parents work; they do so because their peer culture is toxic and because their parents exercise little control over their lives.
Nor does American poverty bear any resemblance to Third World poverty. New York City is awash in welfare programs that confer on the poor benefits that would be unthinkable in rural Mexico or Africa, such as free high-tech medical care, monthly welfare checks, and free or subsidized housing. Family Rewards is an add-on to an already generous safety net; the Mexican version is the safety net. New York’s poor also enjoy a clean, reliable water supply, a public health system that controls environmentally borne infectious disease, a sound transportation infrastructure, and the rule of law. The city’s officially classified “poor” teenagers can be regularly observed sporting the latest designer sneakers.
Nevertheless, lurking beneath the Family Rewards rhetoric about “our impoverished campesinos” was an implicit acknowledgment of a truth rarely spoken in antipoverty circles: it’s the behavior of the inner-city poor that perpetuates poverty, not just “structural inequalities,” rapacious capitalism, or racism. That covert acknowledgment was enough to earn the initiative the opprobrium of many in the traditional poverty industry.
But the program’s proposed cure is potentially worse than the disease: paying families for activities that are part of the normal repertoire of what it means to be a responsible parent or student. (These payments are known as conditional cash transfers, or CCTs.) Randomly selected low-income parents of elementary- and middle-school students in Brooklyn, the Bronx, and Manhattan are paid $25 each month that their child has a 95 percent school attendance record; high school students with a low-income parent in the program receive $50 a month for a similar attendance rate. Elementary- and middle-school students who make progress on annual academic tests net their parents $300 and $350, respectively. High school students get $600 each year that they accumulate 11 course credits (the bare minimum to stay on track to graduate) and another $600 for each New York State Regents exam that they pass. Parents are paid $25 for attending a parent-teacher conference or discussing their child’s test results with a teacher; they receive $50 for getting their child a library card. Taking advantage of taxpayer-subsidized Medicaid services, such as free medical checkups, brings a $200 annual windfall; simply maintaining free Medicaid insurance earns the recipient $20 a month. Working full-time earns an additional $150 a month beyond the existing salary. Seeking education and training while working at least ten hours a week could net a parent $3,000 over three years.
The hubris behind this menu of bribes is breathtaking. Working on the premise that American society didn’t sufficiently reward self-discipline, effort, and achievement, the Family Rewards architects decided that they needed to correct the inadequate signals that the economy and the culture sent to the poor. “The Opportunity NYC–Family Rewards program enabled us to make it worthwhile for families to change their lifestyles to make investments in their futures,” explained New York Deputy Mayor Linda Gibbs at a recent press conference. But of course it’s already worthwhile for families to “make investments in their futures”; the United States still ranks as the world’s primary land of opportunity. The problem is that the poor don’t respond to incentives that are already abundantly present. Nevertheless, convinced of their own superior capacities to engineer sound social signals, the program’s planners arbitrarily made up a schedule of payments that would induce a welfare mother, for example, to make sure that her child went to school every day. Is monthly school attendance worth $25 or $100? Is a single Regents exam worth $600 or $1,200? Ordinarily, markets set prices for true economic exchanges. These were pseudo-economic transactions, however—a fake superstructure imposed on top of noneconomic moral obligations and behaviors that ordinarily bring their own intrinsic reward. Pricing such obligations required a bunch of elite professionals to try to imagine how many shiny baubles they needed to dangle in front of the poor to get them to rouse themselves, a creepy Skinnerian activity demeaning to both the social technicians and their subjects.
The danger with the conditional cash transfers was always that they would work. If children and parents responded to the bribes as the CCT architects intended, the pressures would be enormous to institutionalize the program—currently a pilot involving some 4,800 families, half of whom compose the control group—and to provide public funding for it (the $40 million pilot is currently paid for by the Rockefeller Foundation, the Open Society Institute, and other organizations). Once the expectation of payment for morally and socially responsible actions became widespread, there would be no dislodging it. Poor children would make an effort in school not because their parents had instilled in them the importance of education but because of the expectation of immediate cash. Once word got out that some students in a school were getting paid for their study habits, trying to withhold payments from every child in that school or indeed from every school in the system would be nearly impossible. Poor parents would act responsibly toward their children’s schooling and health not because they understood their obligations as parents, but because certain actions netted an immediate payoff. Any hope that the poor would develop the ability to defer gratification would be destroyed; their already short-term horizons for effort would be shortened further. Society would become divided between a caste that acted responsibly because it understood that it was the right thing to do and another caste that got paid by the responsible caste to follow social norms.
The CCT architects never articulated the answer to another troubling question: if society were not to be permanently divided between socially responsible payers and socially irresponsible payees, what would happen when the bribes ended? Social-science research already suggested an answer: the bribed behavior would promptly trail off. A series of studies since 1971 have shown that once artificial incentives for what should be self-motivated actions are curtailed, the desired behavior shuts down. Third-graders who were given prizes like toys and candy for reading books did not sustain their reading activity once the prizes stopped coming, according to a 2008 study; many stopped reading entirely. A 1999 meta-analysis of 128 psychological studies found that artificial tangible rewards decreased people’s intrinsic motivation and undermined their ability to regulate themselves.
Those who worried about the long-term consequences of the CCTs need not have done so, however. Contrary to the expectations of both its supporters and critics, the Bloomberg experiment had almost no effect on its participants’ behavior. Last month, MDRC (a welfare-research organization that operates the Family Rewards demonstration) published an interim report on the program’s first two years; the findings were stunning. The program had no effect on students’ attendance rates compared with those of students in the control group; it had no effect on average test scores or academic proficiency rates; it had no effect on high school students’ overall accumulation of course credits or successfully passed Regents exams; it had only a negligible effect on the rate at which parents sought a free annual checkup for themselves or their children compared with parents in the control group; and it had a negative effect on the rate at which parents sought education or training for themselves.
While participants reported that their work in mostly unverifiable, often off-the-books jobs increased, their work in jobs covered by unemployment insurance decreased, according to government records. Participants’ efforts to earn rewards declined from year one to year two—even though they presumably understood the mechanics of the program better. Attendance rates dropped for elementary, middle school, and high school students from year one to year two—sometimes sharply—by the same margins as for students in the control group. Fewer high school students tried to take 11 credits or pass a Regents exam from year one to year two.
The only evidence for an across-the board effect on student effort was the increase in the rate at which ninth-graders signed up for 11 credits or sat for a Regents exam. In the first year of the program, 87.8 percent of program ninth-graders attempted 11 course credits, compared with 83.9 percent in the control group of ninth-graders. But the rate at which the program and control-group ninth-graders actually earned 11 course credits was virtually identical: 49.7 percent of the program group and 50 percent of the control group. And by the second year, the difference in rates of signing up for courses and Regents exams narrowed, as did the absolute rate of successful completion. Only 80.5 percent of the ninth-grade program cohort tried to take 11 credits in the second year—a drop of 7.2 percent from the first year—whereas 77.9 percent of the control group attempted 11 credits. As for actual completion of 11 credits, 45.2 percent of the program-group ninth-graders and 45.4 percent of the control-group ninth-graders managed to do so in the second year.
A similar pattern exists for efforts to pass the Regents exam: a slightly higher rate of sitting for a Regents exam on the part of program ninth-graders, but comparable rates of passage. Whether the students in the program group actually studied more to try to earn 11 credits or pass a Regents exam, as opposed to merely signing up in the hope of passing by windfall, is unknown. The answer is important: if students really did make a concerted effort to do their homework and still failed to succeed, that could suggest either a cognitive barrier to learning or a school system so abysmal that it foils student effort.
Despite these lackluster results, the MDRC analysts are not conceding defeat. They present the interim results as a sign that the program is on the path to success, pointing to several bright spots, as they see them. Participants were paid on average more than $6,000 over the first two years of the program, resulting in an 11 percent decrease in the number of families at the federal poverty level compared with the control group. Since one goal of the program was to pump more money into low-income households, the fact that it did so was a positive accomplishment, say the researchers, even though participants’ behavior didn’t change radically. (The program could pay out significant sums despite the relative lack of behavior change because in many cases it paid recipients for activities they were already engaged in, such as attending school or maintaining health insurance.) Program participants increased their use of bank accounts and were less likely to say that they sometimes or often didn’t have enough to eat compared with the control group. Despite transferring large amounts of cash to families, the program did not lessen parents’ overall work level when the increased level of self-reported work is balanced against the decrease in government-recorded work.
The MDRC report cites the complexity of the program as one possible explanation for the lack of more robust results in the first two years. But far from justifying the mediocre outcome, this explanation merely underlines how ludicrous the social-engineering ambitions of Family Rewards were. Many participants couldn’t keep track of the reward structure or of how to claim the money. Some simply ignored the whole concept as too confusing an add-on to their already chaotic lives. There was no uniformity in how parents told their children about the program, note the analysts. Some parents refrained from telling their kids that the parents were getting paid for the children’s attendance and test scores, not wanting to emphasize money as a reason for achievement, according to the MDRC report; other parents discussed the monetary rewards in detail with their children. (The former group of parents display a greater sensitivity to the moral basis of behavior essential for a functioning civil society than the program architects do.) High school students were the target of independent “marketing efforts” by the program engineers, since they received their payments directly, and many “did not fully understand—or believe—that they could earn money for their school attendance and performance,” according to the report. (Again, one applauds the common sense of the participants and wishes that it were shared by the professionals in the welfare industry.)
One mother described to a social worker her own efforts to “market” the program to her high-school-aged daughter, an effort made easier by the existence of separate bank accounts that the program set up for participating teens:
Participant: The more I began to mention the money, I said, “You know, if you do this you’re gonna get this amount. . . .” I gave her the Chase card and I said, “This is your card. . . . The money that’s on it is yours. Whatever you need it for you can use it, but the more you do what you’re supposed to do, the benefits go on this card,” and she was like, “Wow.”
Interviewer: So . . . that was a huge thing for her.
Participant: Oh yeah. Bribery. (Laughter)
Program engineers undoubtedly find this exchange heart-warming rather than horrifying.
One subgroup of students did show positive differences from similar students in the control group: ninth-graders who started the year at an academically proficient level (as measured by the eighth-grade standardized math test). Fifty-one percent of academically proficient ninth-graders who were offered $50 a month to maintain a 95 percent attendance rate did so, compared with 36.2 percent of academically proficient ninth-graders in the control group. (That the paid group of ninth-graders could only muster a 51 percent rate of high attendance, rather than something much closer to 100 percent, is itself remarkable.) As for earning 22 credits over the first two years of the program, 72.7 percent of academically proficient ninth-graders who were offered $1,200 to do so completed the credits, compared with 64.5 percent of the control group. Passing two Regents exams over two years (worth $1,200) was accomplished by 77.6 percent of academically proficient ninth-graders, compared with 71.7 percent in the control group. By contrast, ninth-graders who were not academically proficient—nearly twice as large a group as the academically proficient ninth-graders—showed no difference in performance compared with the control group.
This pattern—more proficient individuals making greater use of the rewards—was repeated throughout the program. The families that earned the most money were the most functional: the parents had higher rates of marriage, full-time work, and education, and lower rates of welfare receipt. What is unknown at this point is whether the families that demonstrated no behavior change made no effort to change or tried to change without effect. If they made no effort to change, is that because they rationally calculated that the rewards were not worth the effort, as the pseudo-economic rhetoric draped over the program would suggest? (Mayor Bloomberg likened the incentivizing effects of conditional cash transfers to private-sector compensation packages in a speech to the Organization for American States in September 2009.) Perhaps, then, the reward amounts were set too low by clueless technocrats, and tripling or quadrupling the payouts would coax a reaction from the buyers still on the sidelines. Maybe $50 a month, say, is simply not that exciting an amount for some of these teens. (One mother whose high school–age son had shown no interest in the program, and whose grades and attendance were slipping, described her son’s access to money: “No, he don’t have a job. He don’t get allowance. I do for him. His sister do for him. His godfather do for him. His cousins do for him. His family do for him. All he does is pick up the phone and say, ‘I want such and such and such,’ and it’ll come through out of the Amazon.com, come to the house.”)
If, however, participants made no effort to change because their lives were so disordered that they could not muster the self-discipline or will to do so, then no amount of technocratic tinkering will change the outcome. Likewise, it could be that some participants actually made an effort to earn rewards but lacked the cognitive wherewithal or self-control to succeed. Either of these explanations seems more plausible than the hypothesis of incorrect pricing, since the least dysfunctional participants in the program collected the most cash.
Two possible policy conclusions follow from the hypothesis that social and cognitive disorganization prevented participants from exploiting the reward structure. The first is: Do nothing. If paying a mother to take her nine-year-old to school every day induces no behavior change in her, it may be time to give up the notion that government programs can erase social and economic stratification. There is a substratum at the bottom of society that will never be raised up by outside intervention. “The premise of conditional cash transfers is that the stresses of poverty cause people to make choices that are not in their long-term interest,” MDRC president Gordon Berlin said at the press conference announcing the interim results. He may have it backward. It is the inability of some people to make choices in their long-term interest that causes poverty, as sociologist Edward Banfield argued four decades ago in The Unheavenly City. The poor have short time horizons, Banfield wrote, the rich, very long ones. No external force can change those psychological dispositions.
One of the few social scientists to have anticipated the results of Opportunity NYC–Family Rewards offers an alternative policy implication. New York University politics professor Larry Mead predicted that the program would have little effect on participants’ behavior. “If the poor were as responsive to [financial] incentives as the policy assumes, they wouldn’t be poor for very long in the first place,” Mead wrote in an e-mail in 2007. If the program had worked, Mead now says, it would have confirmed the notion that the poor are just like us: they respond rationally to opportunity. Instead, the results buttress the message of welfare reform: the poor need strong paternalism and clear moral guidance. Welfare mothers started working not because it was for the first time in their economic interest to do so, but because welfare bureaucrats made it clear that they were expected to work, according to Mead. KIPP charter schools succeed with students because they don’t give them a choice about how to behave. Though Mead finds the improved performance of already proficient ninth-graders in Family Rewards intriguing, he finds more in the study that supports the necessity of wraparound paternalistic intervention. Though the welfare system has exhausted its leverage over welfare mothers once they go to work, the last hope for turning around underclass behavior may be to convert every inner-city school into a KIPP-type academy with minute-by-minute structure and explicit, nonnegotiable rules and expectations.
But the best solution for poverty reduction is the one that is the least likely to pass the lips of liberal policy makers: marriage. As was already abundantly clear before the CCTs, single-parent households are the primary source of long-term poverty in New York City and the country. Looked at from a purely economic standpoint (the least relevant one), a married father provides his children with additional material support and manpower backup when all hell breaks out in a household, as it periodically will. A father also serves as a more credible authority figure than a mother, on average, something that boys particularly need. The recent outbreaks of anarchy in Philadelphia and New York by bands of inner-city youth suggests that the systematic disappearance of fathers from their children’s lives is taking an ever-greater toll on social order.
Family Rewards has cost $33.8 million so far. That amount could have been far better spent on a campaign to educate teenagers and parents about the essential role of fathers. It is by no means clear, of course, that external intervention can change the norms around child-rearing in the inner city. But it is worth a try. At least such a campaign would not undermine fundamental social values.
The conditional cash transfers’ third year will end in September 2010. MDRC analysts are holding out hope that during this final year and the two follow-up observation years, participants’ behavior will show more responsiveness to the program’s reward structure. The city has been promoting its experiment worldwide; the Obama administration has been watching the initiative as a possible basis for future antipoverty programs. If, as seems likely, the final results are as wan as the interim results, much of the poverty industry will claim vindication: the structural inequalities holding back the poor are so massive that no amount of individual self-help can overcome them, the establishment will say. But a more accurate interpretation of the results is that the problems of the multigenerational poor in America stem not from their lack of opportunity, as in the Third World, but from their inability to seize it.