Conditional Cash Penalties in Education

Wisconsin's program that reduced families' welfare payments when their teenagers failed to attend school increased school enrollment and school attendance by more than 3 percent.

In the mid-1980s, Wisconsin began exploring ways to give welfare recipients incentives to work more and to receive a better education. For example, the Learnfare program sought to break the cycle of welfare dependency by giving families greater incentives to monitor teen behavior. When teenagers who were part of this program failed to attend school, their families' welfare payments were reduced by $60 to $190 a month.

In Conditional Cash Penalties in Education: Evidence from the Learnfare Experiment (NBER Working Paper No. 15126), Thomas Dee finds that the Learnfare penalties worked: they increased school enrollment by 3.7 percent and school attendance by 4.5 percent. For students with the highest risk of dropping out, Learnfare increased school enrollment by 25 percent.

Earlier analysis of a ten-county, random-assignment study of Learnfare found that it had "at best modest and short-term effects on its stated enrollment and attendance outcomes." However, Dee concludes that those results were driven largely by the inclusion of data from Milwaukee County, where the random-assignment methodology was poorly implemented. Although eligible participants were supposed to be randomly assigned to either the Learnfare program or regular welfare, in Milwaukee County black teenagers were significantly less likely to be subjected to Learnfare's restrictions than other teenagers.

The Milwaukee County results were complicated by the poor quality of the data systems in Milwaukee County schools. According to a 1995 review, even when sanctions were applied for poor attendance, Milwaukee County took more than twice as long to apply them. When Dee re-examines the data from the Learnfare experiments without the Milwaukee County information, he finds much larger and statistically significant effects of Learnfare on both enrollment and student attendance.

Dee also argues that Learnfare has several unique design features that distinguish it from other programs that provide financial incentives to students. For example, instead of rewarding individual students, it leverages family involvement by sanctioning the family's welfare grant. Furthermore, because it sanctions a grant rather than providing a reward for performance, Learnfare may take advantage of the unique aversion that people typically demonstrate to income losses relative to income gains.

-- Linda Gorman