More lessons from paying people to be less poor
By Barbara Kiviat
Back in 2007, New York City began paying members of some 2,400 poor families to do things like get dental check-ups, open savings accounts, hold down jobs, show up for school, and carry health insurance. Cash incentives were meant to get people with complicated, resource-constrained lives to invest in themselves and their children in ways that would ultimately break the inter-generational cycle of poverty.
The effort, which was inspired by “conditional cash transfer” programs abroad, was the first of its kind in the U.S. Now the program is expanding to Memphis, Tenn., as the mayor there announced yesterday.
Conditional cash transfers (CCTs) have a remarkable ability to bring people together—members of both the left and the right hate the idea. Depending on where you stand, CCTs are offensive because 1) policymakers shouldn’t presume to know what people ought to do, or 2) government shouldn’t pay people to do things they should be doing anyway. I find both sides of that debate disingenuous, unless it’s paired with an argument to end preferential tax treatment of things like home ownership, retirement savings, and student loans—middle- and upper-class equivalents, except for the fact that they are hidden in the tax code and thus distort perceptions of government spending.
I’m much more interested in knowing how CCTs actually change the lives of the poor. The original New York City experiment—and it was an experiment, with thousands of families in a comparison group—saw mixed results. More visits to the dentist, but no change in middle schoolers going to class. A reduction in the use of expensive financial services like check cashing, but a minimal budge in families having health insurance.
The Memphis experiment builds on what was learned in New York City. In Tennessee, the structure of payments is simplified, families may turn to staffers for advice on making plans to earn the payments, and money tied to educational outcomes focuses on high schoolers and adults going back for their GED.
But these are changes to program design, not theory. The same two ideas about how CCTs might transform lives are at play. The first notion—that transfers can reduce material hardship and instability—was clearly demonstrated in New York City, where families earned, on average, $3,000 per year. From a report on program results:
Program group members were less likely to be evicted (2.7 percent versus 4.3 percent for control group members), to have utilities shut off (5.6 percent versus 8.7 percent), and to have their phone disconnected (20 percent versus 25 percent). Program group members reported less food insecurity (not having enough food to eat) and were less likely to report having “insufficient food” at the end of the month (15 percent versus 22 percent). They were also less likely to forgo medical care or fill prescription drugs because they did not have enough money (by 3.9 percentage points and 2.1 percentage points, respectively).
The second notion—that increased stability and resources lead to better long-term planning and goal attainment—is still an open question.
Over the summer, MDRC, the evaluation shop in charge of analyzing the New York City experiment, put out a fascinating report based on interviews with 75 families that participated in the program. The report specifically focused on the education component of the program.
While families often spent program money in ways that directly supported education goals—paying for school supplies, extracurricular activities, even a foreign language trip and a home computer—both adult and high-school-aged participants didn’t often draw a link between these activities and an ability to reach long-term goals. Make no mistake, the goals were there—and long before CCTs came along. Reading the MDRC report makes quite clear that even parents whose families live paycheck to paycheck want their kids to go to college and get good jobs, and that those kids typically share those aspirations.
So then why didn’t families make a connection between their behavior and their ability to reach those goals?
The MDRC report floats a number of possibilities, but the most compelling one is this: families knew they shouldn’t come to count on the money they were earning through the program. The New York City experiment was designed to last three years, and the families participating knew that. In other words, the program inadvertently replicated some of the very instability it was designed to overcome. The result, from the MDRC report, was that:
[T]he program did not tend to inspire hope that families who were experiencing severe poverty would be able to escape from it. This finding is evident in the way that parents and children described their feelings about the end of the program. The desire to maintain a job in a volatile economy, illness or disability, or a desire to stay home with children made changes in work a difficult prospect for parents. As a result, families did not feel that they were able to replace rewards income with work or with a better-paying job after the program ended, and instead talked about the program as an unusual and lucky period in their lives — one in which they would have extra help in making ends meet and would be able to enjoy some greater comforts. (Several parents, in fact, called the program a “blessing.”)
Does that mean a conditional cash transfer program can’t work in the U.S.? Not at all. But it may mean that we’re not going to prove that it can with short-term experiments.