Silvana Krasteva and Huseyin Yildirim have a new and interesting series of results on this question:
Drawing upon the all-pay auction literature, we propose a model of charity competition in which informed giving alone can account for the significant quality heterogeneity across similar charities. Our analysis identifies a negative effect of competition and a positive effect of informed giving on the equilibrium quality of charity. In particular, we show that as the number of charities grows, so does the percentage of charity scams, approaching one in the limit. In light of this and other results, we discuss the need for regulating nonprofit entry and conduct as well as promoting informed giving.
The paper title is “Information, competition, and the quality of charities.” In the basic model, informed donors encourage a “race to the top,” but that competition also consumes excess resources through signaling to get good ratings. An optimum is therefore some mix of informed donors and uninformed donors. As the number of charities becomes very large, however, the chance of attracting informed donors goes to zero, and charity scams end up dominating. According to one cited study, over 90 percent of donors claim to care about quality, but only about 3 percent of them seek out the highest rated charity for a given task.
Under some assumptions, only a single charity provides the public good and all the others end up as scams. Overall I would say that informed donors can be thought of as a scarce resource who cannot be easily leveraged.
For the pointer, I thank the excellent Kevin Lewis.