Do Birds of a Feather Flock Together? Regional Socioeconomic Factors, Agglomeration and Performance of Non-Profit Social Enterprises

Abstract

Nonprofit Social Enterprises (NPSEs) are nonprofit organizations that adopt commercial strategies to sustain their missions. Since nonprofit organizations have traditionally relied on philanthropy and have been known to establish themselves in areas of critical human needs, this shift in the NPSEs’ focus from philanthropy to business creates a timely opportunity to examine potential changes in their location preferences and density dynamics. Using the tenets of organizational ecology, this study examined the context and agglomeration of NPSEs. Specifically, it examined whether socioeconomic factors indicating resource availability are more likely to influence the formation and financial performance of NPSEs in the U.S. than factors indicating resource constraints. The study used county-level government spending, household income and charitable giving as indicators of resource availability, and county-level unemployment rate, poverty rate and ethnic diversity as indicators of resource scarcity. The findings supported a majority of hypotheses, suggesting that NPSEs are less (vs. more) likely to be established in regions where the socioeconomic factors indicate a scarcity (vs. availability) of resources. These findings provide counterintuitive evidence on the purpose and establishment of nonprofit organizations. Additionally, the density of NPSEs had a curvilinear (inverted U-shaped) relationship with their financial performance, indicating that dynamics of competition and density-dependence may be quite influential for nonprofits that adopt a commercial nature. The study primarily contributes to the literature on organizational ecology, nonprofit context and density, and the role of government in the development of social entrepreneurship. It implemented a longitudinal design, and used secondary data from the National Center for Charitable Statistics (NCCS) database and the U.S. census survey for analysis. The hypotheses were tested using multilevel modeling (MLM) and polynomial regression.