Working Paper: CEPR ID: DP9941
Authors: Dean S. Karlan; Daniel H. Wood
Abstract: We test how donors respond to new information about a charity?s effectiveness. Freedom from Hunger implemented a test of its direct marketing solicitations, varying letters by whether they include a discussion of their program?s impact as measured by scientific research. The base script, used for both treatment and control, included a standard qualitative story about an individual beneficiary. Adding scientific impact information has no effect on whether someone donates, or how much, in the full sample. However, we find that amongst recent prior donors (those we posit more likely to open the mail and thus notice the treatment), large prior donors increase the likelihood of giving in response to information on aid effectiveness, whereas small prior donors decrease their giving. We motivate the analysis and experiment with a theoretical model that highlights two predictions. First, larger gift amounts, holding education and income constant, is a proxy for altruism giving (as it is associated with giving more to fewer charities) versus warm glow giving (giving less to more charities). Second, those motivated by altruism will respond positively to appeals based on evidence, whereas those motivated by warm glow may respond negatively to appeals based on evidence as it turns off the emotional trigger for giving, or highlights uncertainty in aid effectiveness.
Keywords: aid effectiveness; charitable fundraising; pure altruism; warm glow
JEL Codes: D64; H41; L31; O12
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Prior donation amounts and demographics (D64) | Control for confounding factors in donation behavior (D64) |
| Larger prior donors presented with effectiveness information (F35) | Increase in likelihood of giving (D64) |
| Small prior donors presented with effectiveness information (D64) | Decrease in giving (D64) |
| Average donation behavior for the full sample (D64) | Masking significant heterogeneity between large and small donors (D64) |