Donor coordination in project funding: Evidence from a threshold public goods experiment

Highlights

• On crowdfunding sites or in charitable giving, multiple projects vie for limited funding.

• We experimentally study how the presence of multiple options affects contributions.

• The presence of multiple projects decreases coordination and total contributions.

• Coordination is not a problem when one project stands out on its merits.

• Important for crowdfunding sites to highlight the best projects

Abstract

We present results from an experiment with multiple public goods, where each good produces benefits only if total contributions to it reach a minimum threshold. The presence of multiple public goods makes coordination among participants more difficult, discouraging donor participation and decreasing the likelihood of any public good being effectively funded. Applied to the case of fundraising, the results show how overall donations and the number of effectively funded projects may both decrease as the total number of projects vying for funding increases. The analysis considers whether making one of the contribution options salient, either through its merits or by arbitrarily choosing one to feature during the experiment, helps overcome the increased coordination problem. The results have implications for the growing popularity of crowdfunding websites, and suggest benefits to these sites from helping donors compare and identify the most promising projects.

Introduction

Fundraising can make the difference between success and failure for a new non-profit or business venture. Although attracting funding is no guarantee of success, failing to attract funding often guarantees failure. Sometimes, an organization attracts a single individual, who alone provides enough capital to bring the organization through the early stages of development, allowing it to establish itself as a viable venture. Other organizations rely on smaller donations from multiple individuals to raise the level of capital necessary to become a viable, successful project. Such “crowdfunding” has become more prominent in recent years as the internet (including websites Kickstarter.com, IndieGoGo.com and Kiva.org) has made it easier for projects to reach out to a broader set of smaller donors in their fundraising efforts. The popularity of these crowdfunding sites has led to a surge in the number of projects seeking funding from the public.

In August 2013, there were 3957 separate projects vying for funding on Kickstarter.com (Lau, 2013). This number continues to grow, with 6802 active projects vying for funding in October 2014. These projects range from charity initiatives designed to help a community, to art exhibitions and films, to the production of innovative consumer products. Potential financial backers visit the website and can select projects to which to pledge funding. A project receives its funding only if the total amount pledged reaches a threshold amount. IndieGoGo.com, which follows a similar model, had 4348 projects seeking funding in August 2013 (Lau, 2013). In a related way, Kiva.org brings together individuals to provide micro-finance loans for small business ventures around the world. Projects remain on the site for up to 30 days, and Kiva passes payments along only to those projects for which pledged contributions reach 100% of their funding goal. We counted 5131 separate projects actively seeking funding on the site.

Although some of the projects on these sites stand out due to a particularly creative project design or some other feature that grabs public attention, most of the projects are similar to a number of other initiatives simultaneously seeking funding. For example, on Kiva.org, there are 40 projects seeking less than $350 in funding to purchase clothes for one-person retail clothing businesses in developing countries. Many of these projects are essentially indistinguishable from one another, with the borrowers asking for similar funds, for similar purposes, and providing similar backstories and information on the site.

The large number of projects simultaneously seeking funding raises a variety of concerns. First, it can lead to an inefficient distribution of donations across projects. Successfully funding a project typically requires contributions from multiple donors. As the number of alternative projects increases, coordination among donors becomes increasingly difficult. This increases the likelihood that a donor pledges her donation to a project which eventually fails, an inefficient allocation of funds. Second, the number of projects can discourage donors from giving in the first place. This may result if fully-rational donors recognize the difficulty in coordinating, and choose not to give due to lower expected payoffs. It may also result if donors suffer from the paradox of choice, which suggests that people are less likely to participate as the number of options to choose between increases, even in the absence of coordination problems (e.g. Schwartz, 2004). Together, these concerns suggest that the proliferation of giving options will decrease contributions, and result in the less-efficient allocation of the remaining contributions.

This is a real concern on crowdfunding sites where not all beneficial projects reach their fundraising goals. The portion of listed projects that eventually fail to receive funding was approximately 56% on Kickstarter.com, and 33% on IndieGoGo.com in August 2013 (Lau, 2013). By our own estimates using Kiva.org data, 8% of projects on Kiva.org failed to receive funding.¹ Although this represents a much higher success rate than the other sites, failure to attract funding remains a sizable concern.

Our paper explores how the presence of multiple giving options may affect contributions and fundraising success. We conduct a laboratory experiment involving donors contributing across multiple threshold public goods, which represent projects vying for funding. The analysis compares two treatments: in one, donors choose how much of their endowment to contribute to a single good; in another, donors choose how much of their endowment to contribute to each of four similar goods. Because the four goods in the multiple option treatment are ex ante indistinguishable, there does not exist any coordination device to guide donor choice, clearly increasing how difficult it is for donors to achieve coordination. The increased difficulty in coordination in turn may decrease incentives that donors have to contribute and the probability of project success. Consistent with theoretical predictions, we find significant evidence that going from one to four contribution options significantly decreases the coordination rate and total contributions across all projects. That is, increasing the number of projects vying for funding not only decreases the probability that any given project succeeds, it also decreases overall contributions and the expected number of successful projects.

In addition to highlighting potential costs of multiplicity using a treatment with four similar public goods, the analysis explores how the results change when one of the goods stands out compared to the alternatives. In a crowdfunding setting, this may be the case if one of the projects stands out on its merits as being more promising than others. It may also be the case if the expected merits of all projects are identical, but one of the projects is arbitrarily highlighted or featured on the fundraising website or by the media. To analyze these possibilities, we consider two additional treatments, in each of which three of the available public goods are identical, and one additional public good is made salient. Depending on the treatment, the salient good may stand out due to it providing higher potential payments than the other three alternatives, or due to it being featured on the experimental computer screen at the time the donors make a contribution decision.

In the treatment in which one of the options stands out on its merits as offering a higher potential benefit, we find that the contribution pattern is almost identical to the case in which there is a single contribution option. Increasing the number of contribution options does not increase the donor coordination problem, does not decrease overall funding, and does not decrease the probability that the efficient option succeeds when the additional options are clearly less efficient than the original. When donors can identify the most promising contribution options, the coordination problem that arises from the presence of multiple options disappears.

In the treatment in which all contribution options are identical, and a non-merit-based signal directs donor attention towards one of them, contributions tend to be higher than in the case with four indistinguishable options, but lower than in the case with a single option or where one of the options stands out as most efficient. Although the differences between contributions in the random signal treatment and the other treatments tends not to be significant, the pattern of contributions consistently suggests that randomly featuring one of the contribution options can help reduce the coordination problem, but does not fully overcome it. Importantly, these results suggest that fundraising sites may be able to alleviate some of the coordination problem by directing donor attention to certain projects, even if such projects are selected randomly from a set of otherwise indistinguishable options. When possible, however, the coordination problem is more effectively overcome when donors can identify the most promising projects. This suggests the importance of individual fundraisers emphasizing the merits of their projects and fundraising websites or organizations featuring projects based on their merits.

Overall, our results show that salience can help overcome coordination problems and lead to greater efficiency in contributions. This is not to say that making one of the options salient is sufficient to facilitate donor coordination. In an additional treatment, we consider the case in which one contribution option stands out as salient because it is less efficient than three indistinguishable alternatives. Importantly, the less efficient good in this treatment is still efficient enough relative to the alternatives that it would be advantageous for donors as a group to focus on the salient option, where uniqueness makes successful coordination relatively easy to achieve. Despite this, the group tends to ignore the salient option and instead direct contributions to options which offer moderately higher payoffs, but where coordination is significantly more difficult to achieve (due to the multiplicity of indistinguishable options). This suggests that donors as a group may focus on potential payoffs rather than expected payoffs when choosing where to donate. Salience fails as a focal point when the salient option does not also offer one of the maximum potential payoffs, even when payoffs are such that the group would be better off focusing on salience.

The majority of the analysis focuses on treatments in which donor groups cannot afford to fully fund all four contribution options. This is based on the general view in fundraising that there is competition for limited donor contributions. As Walter Sczudlo of the Association for Fundraising Professionals explains for the case of charitable giving, “[the] proliferation of charities is creating a huge competition for donor dollars. There are so many charities now going after so few dollars and it's getting parsed out so finely.”² In alternative settings, however, limited availability of funds may not be an issue. With this in mind, we conducted another experimental treatment in which donor groups can collectively afford to fund all four public goods, and individual donors can afford to fund entire public goods themselves. In order to compare contributions across treatments, we effectively weaken the budget constraint by reducing the relative thresholds and payoffs associated with the four public goods, while keeping endowments constant. This is similar to splitting the public good in the one-good treatment into four independent sub-projects.³

In this alternative treatment, weakening the budget constraints effectively makes it easier to achieve the threshold on any given public good, while at the same time increasing concerns about free riding. We find no evidence that weakening the budget constraint leads to higher contributions. If anything, the opposite is true, with total contributions being lower in this alternative treatment than when the group can fund at most one good.

Taken together, our results show that much of the coordination problem introduced by multiplicity may be eliminated if donor attention is directed towards a single contribution option. However, for this to most effectively facilitate coordination, the highlighted option needs to be perceived by donors as the most promising option. This will be the case if fundraising sites feature projects based on their merits. Randomly selecting options to feature does less to solve the coordination problem, and does nothing in the event that the featured option looks less promising than alternatives.