Association for Public Policy Analysis & Management

Despite the surge of academic interest in cross-sector collaboration in the past two decades, the substantive focus of public-nonprofit collaboration scholarship tends to be predominantly focused on social service industries and the contracting regime where a single direction of funding flows from government to nonprofit organizations is assumed. As a result of this single funding flow assumption, theory-building in government-nonprofit relationships is largely biased toward the perfect distinction between provision and production. Nonprofit and for-profit providers are regarded as an alternative arrangement of public service production while public service provision is still decided and financed by government agencies (McGinnis 1999). Empirical studies in government-nonprofit collaboration also tends to focus more on the influence of government funding on nonprofit finances and mission (Guo 2007; Brooks 2000). However, when there is a significant reverse or two-way funding flow from nonprofits to government agencies, can we flip the coin and ask: what is the impact of philanthropic funding and cross-sector collaboration on the capacity of local governments?

Research has produced a multi-dimensional taxonomy of the activities conducted by park-supporting charities (Gazley, Cheng, & Lafontant 2015). These nonprofits are not only engaged in service production activities (parks maintenance), but also in service provision activities (advocacy and planning). An important next question is how those park-supporting nonprofits impact the capacity of local parks and recreation departments? It is possible that both strengthening and weakening pressures occur, since these organizations can generate more public enthusiasm towards parks and therefore increase the public budgets for parks and recreation, but they could also encourage city governments to reduce their level of public spending on those services.    

Data come first from a unique national and multi-year database of local park-supporting nonprofits. Each case has substantial coded characteristics (revenues, activities, and community characteristics). A second data source is the annual city park facts report issued by the Trust for Public Land, which includes comprehensive data on major U.S. city park systems from 2009 to 2015. Because of the existence of this simultaneous and two way causality existed in our key independent (nonprofit spending) and dependent variables (parks budget), propensity score matching method will be used to model this complex cross-sector funding interactions. By using multiple waves of data and the propensity score matching method, the first wave of data will be used to calculate the propensity for cities to have a certain level of nonprofit spending on parks and recreation. The propensity score and the second wave of data will be used to estimate the level of public spending in the subsequent period of time. By controlling for the propensity of having a certain level of nonprofit support, the problem of reverse causality will be alleviated. This modeling approach also presents a better way of capturing the dynamic interactions and feedback loops presented by multiple collaborative governance frameworks (Ostrom 2011; Emerson et al. 2012).