Do charities get more when they ask more often? Evidence from a unique field experiment

- The net effect of a single request to donate to charity is not easily observable, certainly not when multiple charities address the same individuals.
- We alleviate these observational difficulties by carrying out a field experiment in which five large charities cooperate.
- With the unique data that we collect, we study the impact of sending more requests on total donations.
- We find that there is a negative competitive effect on requests from other charities, but this effect dies out rapidly. Soon after the mailing has been sent, it is only a strong cannibalization of the charity's own revenues that prevails.
- Our conclusion is thus that not much coordination across charities is needed to increase revenues.

Charitable organizations send out large volumes of direct mailings, soliciting for money in support of many good causes. Without any request, donations are rarely made, and it is well known that each request for money by a charity likely generates at least some revenues. Whether a single request from a charity increases the total amount donated by an individual is however unknown. Indeed, a response to one request can hurt responses to others. The net effect is therefore not easily observable, certainly not when multiple charities address the same individuals.In this paper we alleviate these observational difficulties by carrying out a field experiment in which five large charities cooperate. With the unique data that we collect, we study the impact of sending more requests on total donations.The results indicate that there is a negative competitive effect on requests from other charities, but this effect dies out rapidly. Soon after the mailing has been sent, it is only a strong cannibalization of the charity's own revenues that prevails. This empirical finding suggests the important conclusion that not much coordination across charities is needed to increase revenues. We also demonstrate that charities need sophisticated evaluation tools that do not ignore the effects of cannibalization.