Financial Incentives Program Pays Off 3 Years After Poor Results

Three years after the cash-for-good-behavior experiment in the Bronx was deemed a costly flop, it quietly continued to produce impressive positive changes in adult and child behavior. Experts initially dismissed the proposal to pay members of low-income families in the Bronx borough of New York City for things like keeping medical appointments, passing classes and remaining employed, but after scientists tightened the feedback loop between good actions and rewards, they managed not only to get people to alter their behaviors, but they were able to make those changes stick.

Among the changes adopted for the second phase of the experiments is shrinking the period between the action triggering a financial incentive and the actual payments. Instead of paying students at the end of the year for passing their classes, for example, they were instead paid monthly for maintaining good grades.

“We really wanted to tighten up that loop so that it would be a much more direct reward, closer to when the activity had taken place,” said Kristin Morse, executive director of the city’s Center for Economic Opportunity, which created the program.

The experiment is part of a broader effort to answer elusive questions in economics: Can you pay people to change their lifestyles, and if so, how much cash does it take?

“We are not that good at changing behavior using incentives,” said Uri Gneezy, a professor of economics and strategy at the University of California-San Diego. “The reason is not that incentives cannot work, but that we don’t know how to use them yet. We are improving, but we are not there yet.”

The experiment was initially a brainchild of New York City Mayor Michael Bloomberg, who believed that private sector strategies like financial incentives could be useful for encouraging “good behavior.” Lisa Fleisher of The Wall Street Journal reports that the program was a joint effort between the city, research firm MDRC and local community groups. Since its launch, it enrolled more than 4,800 families and has so far cost $20 million of the $30 million in total funding raised.

Critics latched on to the program early on, pointing out that incentives were a poor way to encourage good choices in cases when circumstances made such choices impossible. Could one expect money for continued employment to encourage people to stay in jobs in an area where no jobs were available and involuntary unemployment was endemic?

The full study showed little effect on doctor’s visits, employment, and elementary and middle-schoolers’ test scores or attendance. Some of the encouraging initial results diminished over time. Families had more money in the bank, but that was attributed to the program’s cash payments.

There were bright spots: Students already proficient in reading were more likely to graduate. Nearly 75% of students who were ninth-graders when the program launched graduated on time, compared with 67% of the control group. Researchers attributed the increase to the payments, which included $600 for passing each Regents test required for graduation.