About a year ago on the Evidence Base, I reported on a study in which we asked people whether they would make charitable contributions to different food aid programs. We wanted to know whether there are differences in willingness to donate to programs in which food aid is given for free, versus programs in which recipients have to pay part of the cost. We found that people were more likely to donate to programs in which recipients have to pay a small part of the cost. This lends support to the economics theory of screening, which proposes that asking recipients to pay part of the cost screens for people who most need and value the aid. We studied donation behavior using a “revealed preference” approach, which is a favorite of experimental economists as it replicates real-life situations and asks people to make actual decisions (for instance, donate money to a charity), rather than asking people directly what they would do.
However, when we presented the study, other academics pointed out to us that it would be a good idea to ask direct survey questions about what types of aid programs people prefer and why. So my co-author Sally Sadoff (UC San Diego) and I decided to go back to the field and survey people about their views on and preferences for different aid programs. We again partnered with the Understanding America Study (UAS) internet panel. This panel is unique in that our respondents were a group of nearly 5,000 Americans representative of the U.S. population.
In the new survey, we described a program that provides healthy food to low-income families. We asked people which program they would support: one in which the food is free, one in which families who receive the food pay $1 or $5 of the cost, or one in which families who receive the food have to spend 5, 25, or 45 minutes registering for the program.
The figure below shows the percentage of respondents who preferred each type of program. We see that the most popular program was when the food aid was free, followed by the program required families to pay a small cost ($1 or 5 minutes of their time). This result differs from our earlier study, where we found that people donated more to programs in which families had to pay a small cost. Hence, we might tentatively conclude that what people actually do (i.e., what they donate to) differs from what they say (i.e., what they prefer in the survey). However, the conclusion is tentative since the two studies cannot be directly compared for a variety of reasons.
We also found that the least popular were programs that required an onerous registration process (25 or 45 minutes) or that required families to pay $5. This is in line with our prior study: people don’t donate to, and don’t say they like, programs that place a large burden on recipients.
Next, we asked respondents whether they agree or disagree with common reasons for supporting each program. For example, we asked whether people agreed that having families pay for the program is a good way to identify those who most value and need healthy food. About 40-50 percent of people agreed or strongly agreed with this and similar types of screening statements. About 50 percent agreed or strongly agreed that having families pay part of the cost also allows programs to serve more people, while 50 percent also said that they worry poor families may not be able to afford the cost.
Overall, we observed that there is heterogeneity in preferences, with some people preferring free programs and others preferring programs that require families to incur some costs. This mirrors a similar split in policy and practice: for example, some U.S. policymakers advocate for aid programs that require recipient work requirements, while others are against such a policy. And, some charities provide aid free of charge, while others require recipients to work for aid.