April 28, 2010
In March, MDRC released its initial evaluation of New York City’s Family Reward’s demonstration program. Family Rewards provides cash payments to very low-income parents (below 130 percent of the current poverty line) who meet various conditions related to their children’s education (including children’s school attendance and scoring above a certain level on standardized tests), family health (maintaining health insurance coverage and preventive care), and their own employment and training (full-time work and participating in certain education and training activities). Family Rewards is based on “conditional cash transfer” programs that currently operate in Latin America, particularly Mexico’s Oportunidades program.
The headlines from press accounts of the evaluation (NYT: “New York to End Program to Give Cash to the Poor”; AP: “Money for Good Habits Doesn’t Change Lives”) frame the demonstration as having little impact, but the actual results are more nuanced. Most notably, families enrolled in the program experienced average monthly income gains of $366 compared to families in a control group. Family Rewards payments amounted to about 80 percent of the difference. Families in the demonstration also experienced substantial reductions in various economic hardships and financial strain. They were less likely to have their phone and utilities disconnected or be food insecure, and more likely to report an improved financial situation compared to the previous year.
The results related to children’s education, however, were quite limited. There was generally no improvement in the test scores or school attendance of children in the program (although attendance was already quite high). One exception was for a subgroup of high school students: those students who scored at or above proficiency level in the previous year (before entering the program) saw some gains in attendance and likelihood of moving forward to the next grade. One caveat to these results is that they are based on only one year of data. It may be that additional time is needed for the program to have an impact on educational outcomes; we’ll know in 2013 when the final longer-term evaluation is released.
What I take away from the evaluation is that providing income supplements to low-income families increases their income and reduces their levels of economic hardship. This may seem like an insignificant point, but remember that conservatives and even many mainstream anti-poverty researchers believe that public income supplements are counter-productive because they have a large “work disincentive” effect unless conditioned explicitly on employment. That wasn’t the case here. (Although Family Rewards isn’t a clear-cut test since one of the 22 incentives was for maintaining full-time employment.)
The big question with Family Rewards is whether it’s a better way to deliver income supplements to low-income families than other alternatives. The Bloomberg administration describes Family Rewards as the “first major conditional cash transfer initiative” implemented in the United States. This is true in a narrow programmatic sense—the program is the first U.S. demonstration directly modeled on the Latin American CCT programs—but it’s not really the case in the broader conceptual sense implied by the term conditional cash transfer. Federal and state governments provide all sorts of transfers that are conditioned on meeting various requirements. Federal tax law, for example, provides subsidies conditioned on homeownership (the mortgage interest deduction), on earnings (the Earned Income Tax Credit), and education (the Hope Tax Credit, Lifetime Learning Credit and others).
While Family Rewards shares these programs’ conditionality, it’s different in other ways: 1) the transfers are not explicitly linked to costs associated with meeting the conditions; 2) the program is limited to very-low income families; 3) it provides a long list of micro-incentives under a single programmatic umbrella administered by a non-profit agency. This approach might make sense in Mexico and other middle- and lower-income countries where existing means-tested social insurance programs are limited or non-existent, but I’m skeptical that it makes much sense in the United States where we already have established structures for delivering means-tested social insurance and conditional transfers.
So, instead of 22 incentives bundled in a single program, I’d opt for focusing on a handful of benefits that could be made progressively universal and are designed to offset specific costs. To a large extent, this would simply mean extending existing benefits to low-income families. For example, the Lifetime Learning Tax Credit offsets the costs of post-secondary education and training for adults, but it’s only available if you have federal income tax liability. The Earned Income Tax Credit provides benefits conditioned on earnings, but the benefits are teensy for low-wage workers without children and youth are completely excluded. Both credits should be made more inclusive and progressive. In terms of new benefits, I’d focus on something for youth in the 16-24 age range, a group who are largely ignored by current social policy (and left mainly to the criminal justice and educational systems). One possibility would be a credit for completing high school (in part to explicitly reduce economic pressures to leave school for work) and, after doing so, moving into work or post-secondary educational and training.
