Learning by fund-raising

From experience, fund-raisers become more efficient solicitors. This paper offers a full characterization of the optimal solicitation strategy when there are learning economies. The characterization is recursive, and does not require equilibrium computations. Our main results are that (1) the fund-raiser may now solicit a " net free-rider"-a donor whose contribution falls below the marginal solicitation cost; (2) the number of solicitations is non-monotonic in the learning rate; and (3) a solicitation technology with high overhead but low marginal costs is likely to be optimal if income distribution in the population is relatively homogenous. The complementary case with a decreasing returns to scale technology is also considered.