The effect of recommended contributions in the voluntary provision of public goods

Abstract

MELANIE MARKS [*] Recommended contributions are often observed in fundraising campaigns for charitable and other public goods. We present an experiment investigating the impact of recommended contributions in a voluntary threshold public goods process. We find that when valuations for the public good are heterogeneous, recommended contributions significantly increase the likelihood of efficient provision, although when valuations are homogeneous, the effect of recommendations is less compelling. This article represents a first step in understanding recommended contributions and other nonbinding, cheap-talk announcements in public goods provision and charitable contributions. (JEL H41, C72) I. INTRODUCTION Recommended contributions are often observed in fundraising campaigns for charitable and other public goods, but very little previous work has investigated their effectiveness. This article is the first in a stream of research designed to investigate the impact of recommended contributions on the provision of public goods. In this study we experimentally examine the impact of a nonbinding recommendation of how much each giver should contribute toward a threshold (binary) public good. We find that recommending contributions improves voluntary provision of the public good when the contributors value the good differently. In contrast, recommended contributions are less effective when valuations for the public good are homogeneous. Since the former case is more realistic than the latter, these results provide limited support for the likely effectiveness of recommended contributions in this setting. Field studies from the marketing and psychology literatures investigate the impact of recommended contributions on charitable appeals more generally, with mixed results. Weyant and Smith (1987) used a mail campaign of the American Cancer Society; they found no significant differences in either probability of contributing or amount contributed when recommended contributions of $5-25 or $50-250 are included in the mailing. In contrast, Fraser et al. [1988] found that recommending a $20 donation for the Capitol Area Humane Society led to no differences in the probability of donation but a significantly higher amount donated than no recommendation. Brockner et al. (1984) also found differences between charitable contributions to the National Reye's Syndrome Foundation with and without recommended contributions of either $1 or $5. In their work, the addition of either recommended contribution increased the probability of a contribution but did not change the amount contributed. [1] One concern with this research is that the valuation of the public good by potential donors is not known to the experimenter. Thus, it is hard to determine why the recommendation is having the effect it has. For example, the recommendation could be influencing the contributors' belief of what others are contributing, causing higher contributions via reciprocity, as in Croson (1999), or coordination, as in Dale and Morgan (1999). Alternately, it could be changing the donors' beliefs about the value of the public good being provided by sending a signal about its quality, as in Vesterlund (1998), or about the prestige of the giver, as in Harbaugh (1998a, 1998b). In the experiment reported here, we use induced valuation to control the benefit from the public good to the subjects, as described in Smith (1982). Thus, we control the value of the public good to the donor and eliminate a class of explanation involving the recommendation changing it. Our experiment examines voluntary contributions to a threshold public good as introduced in Bagnoli and Lipman (1989). In this mechanism, contributions for the public good are solicited, and if the sum of the contributions meets or exceeds the cost threshold, it is provided. Concrete examples include soliciting money for a group to hire a lobbyist, described in Bagnoli and McKee (1991), and a power company to build an energy-efficient power plant, described in Schulze (1995). …