Association for Public Policy Analysis & Management

There is wide geospatial variation in access to the nonprofit human services safety net. My paper “Holes in the Safety Net” shows a statistically significant relationship between rurality and lower levels of human services spending nationally, even after controlling for other county and state level characteristics. In analyzing the data I realized that there appear to be quite a few counties where there is no human services spending at all. In 2013, there were 360 counties nationally that showed no nonprofits administering social services in the IRS 990 data. Several states, including Georgia, Kansas, Kentucky, Mississippi, Montana, Nebraska, North Dakota, South Dakota, Texas, Utah, and Virginia, account for 273 of the 360 counties of interest.

It is possible that counties that appear to offer no human services are a function of the shortcomings of the IRS 990 data, rather than places where no services are offered. Because it relies on organizations’ tax returns, the 990 data fails to capture some provision of human services, including very small providers, religious providers, and services provided by county governments. On the other hand, it is possible that the dataset is showing an important indicator of places where there is very little or no human services provision at all. The mental health provider shortage data suggests that 641 counties nationally did not have a single social worker in 2015 (Larson, Patterson, Garberson, & Andrilla, 2016). The social worker provider shortage numbers, combined with the lack of providers indicated in the 990 data, suggest that it is also possible that many of these counties provide no human services at all.

This paper will investigate the question: are the counties that do not show human services provision in the 990 data truly not providing any human services, or are some service provision not being captured by the 990 data? The paper will also describe the characteristics of the possible human services desert counties and map which services are offered in them. This project uses IRS 990 data on social services funding. The dataset originates from the National Center for Charitable Statistics, which includes the total annual expenditures of all nonprofit human services organizations that file an IRS-990 form.

The paper will use a survey instrument to survey the possible human services desert counties by phone to assess whether there is actually no funding for human services available in those counties, or whether services are being provided by providers that are not captured in the 990 dataset, such as very small providers, religious providers, or county government. The survey will ask about individual types of human services to attempt to capture both the service providers and the types of services offered. Once completed, this project will provide a picture of the number of counties nationally where there are no or very few nonprofit human services being offered. This will further policymakers’ understanding of spatial access to the social safety net and the characteristics of places where the safety net may barely exist at all.