Panel Paper:
The Structure of Nonprofit Competition and Overhead Spending Ratios
*Names in bold indicate Presenter
Three major challenges exist for research on the effects of competitiveness in the nonprofit sector (Wong 2005). The first challenge is defining the market. Neither field of interest nor geography is necessarily an appropriate market delineation for nonprofits, although this is a required part of traditional measures of competition like the Herfindahl-Hirschman Index. The second challenge is measuring intensity within a market. The number of nonprofits or even their sizes may explain little about competitive intensity since nonprofits can collaborate or draw on distinct donor groups to avoid competition. Finally, many measures of competition generate endogeneity concerns. For instance, measures where the market is determined by the geographic scope of donors will result in larger scopes for organizations with higher unobserved quality (Kessler and McClellan 2000).
To overcome these challenges and examine the effect of competition on overhead, I use individual-level giving data from the Combined Federal Campaign (CFC), which is the workplace giving program of the federal government. This unusual data set allows me to observe pledges (nonprofits selected and amount donated to each) by federal donors between 2008 and 2013. The data covers 59 local markets, approximately 850,000 pledges, and 22,500 public charities. Using social network analysis techniques, I empirically construct the network of nonprofit-donor ties and measure the degree of connectedness for each observed nonprofit—the extent to which a nonprofit shares its donors with other organizations. The donor data provides an empirical definition of the market, while the connectedness measure quantifies intensity. I measure overhead spending using the administrative and fundraising expense rate calculated by the CFC. I regress overhead rate on my connectedness measure of competition, total dollars given to each nonprofit (which acts as a proxy for unobserved quality), and other relevant controls (subsector and year fixed effects) to understand the relationship between competition and overhead spending. To evaluate the relative strength of this approach, I compare the calculated levels of competition from the network data with a Herfindahl-Hirschman Index calculated using sector and geographic information from the IRS Form 990.