Rewarding Good Behavior

March 31, 2007
Posted by Jay Livingston

Many people think of welfare as helping the unfortunate. In this view, the government should ensure that everyone, even the poorest members of society, can feed and shelter their children. Yet for as long as I can remember, conservatives have hammered on welfare as “rewarding bad behavior” — if you give women these princely sums (a couple of hundred dollars a month) when they have babies, of course they’ll go and have more babies whether they’re married or not.

The evidence doesn’t support this idea, but the moralistic and punitive mentality remains unchanged. The most striking feature of the “welfare reform” of the 1990s was this punitive approach: if you don’t get a job, we’ll punish you by taking away your welfare grants.

Now, a new program in New York City keeps the same conservative view, but instead of punishing bad behavior, it seeks to reward good behavior.

The city will select an experimental group of 2500 poor families and give them cash for good behavior in

• education (children’s attendance and performance, parents’ involvement)
• health (regular check-ups at doctors and dentists)
• work (job training, looking for work, working)

The total payments, if a family does everything right could add up to $5000 for a year.

The official announcement was full of optimism, and the press coverage has been favorable. I like the idea that it’s based on reward rather than punishment. At a time when the very richest are getting astronomically richer, it just doesn’t seem right to have government taking money away from the very poorest. I also like the built-in experimental design — a study will also follow a control group of 2500 similar families who are not offered the incentive. (As Mike Kellerman at Harvard’s Social Science Statistics blog notes, “The image of 2,500 families randomly selected to not receive benefits probably doesn't do much to help the cause . . .”

But I have a couple of reservations. First, I worry that the rewards won’t work because they are too small and too long range. Suppose we offer a reward of $50 or even $100 for good school attendance. If the family can’t get the cash till the end of the term, will that far-off promise be enough to offset whatever pressures there are in the present that have been pulling the kids away from school? These rewards are based on the assumption that the poor have the luxury of a longer-term perspective. That perspective is much more available in the comfort of a middle-class income than in the world of poverty with its frequent financial crises.

Even middle-class people may not act “rationally” to maximize their rewards. In my classes, I offer bonus points for papers turned in on time. On most assignments, about half the students don't take advantage of the offer. Other things in life get in the way of doing the paper on time.

Suppose that the study does prove a success, and that the experimental group does better in those three areas. Is the program an answer to the problem of poverty?

Programs like this one assume that poverty is matter of individual characteristics; they ignore the role of the broader economy. Welfare rolls were reduced in the 1990s, and Clinton’s welfare reform may have played a part. But so did a booming economy, which was certainly not caused by welfare reform. In those full-employment years, employers were paying living wages to people whose personal characteristics might have caused employers and economists a few years earlier to write them off as “unemployable.” In a weak economy, all a program like this might be expected to do is to convert the welfare poor into the working poor.