Community Development Block Grants: Exploring Organization and Neighborhood Networks

Background & Purpose: Multiple studies suggest that reaching a critical mass of spatially targeted resources using Community Development Block Grants (CDBG) improves conditions in targeted neighborhoods compared to neighborhoods that do not receive funds. This study, which is the first to use network analysis in this context, examined CDBG spending in a mid-sized, Midwestern city. The purpose was to examine networks of targeted and city-wide CDBG spending, understand how neighborhoods of varying poverty levels receive projects, and explore how neighborhoods form networked communities based on shared projects.

Methods: This study used network analysis to examine ties between CDBG funded projects, represented by the implementing organization (n=64), and neighborhoods, defined as census tracts (n=106). Data were extracted from administrative documents related to CDBG project funding for the study site as well as 2015 American Community Survey estimates. Data consisted of project details, affiliated organizations, funding, location, neighborhood poverty rate (%), and neighborhood poverty category (1 “<20%” to 4 “>40%”). In this network, nodes represent organizations and neighborhoods, with the ties indicating spatially targeted projects. Both network visualizations and descriptive measures were examined for the organization and neighborhood networks. Additionally, simple linear regression was performed to examine the relationship between network connections and poverty rates for neighborhoods. Finally, community detection was conducted on the neighborhood network to identify important neighborhood subgroups.

Results: The full network consisted of one large component and many isolates, which represented either city-wide projects or neighborhoods not targeted by a project (density=.01). Visualizations of the independent networks for organizations and neighborhoods indicated that the highest levels of funding are allocated to city-wide projects or administration, and although there are some high-poverty neighborhoods that are not targeted directly by projects, most are included. On average, each neighborhood is targeted by 1 (mean= 1.47) organization, and each organization serves about 2 neighborhoods (mean=2.44). Twenty-two neighborhoods have a high number of organizations implementing projects (>2), most of which are high poverty areas (mean = 39.07, SD=13.70). Regression analysis indicated a significant positive association between poverty rate and number of projects (b= .04, se=.007). Subgroup analyses indicated that neighborhood subgroups were not determined by poverty level.

Conclusions & Implications: Network visualization indicates that most CDBG spending is distributed widely, either through administration or city-wide projects, with an average of 1 targeted project per neighborhood. Previous research suggests that this wide distribution of project funds may not be as effective as targeted distribution. Despite this, the findings indicate that high-poverty neighborhoods are well-represented in terms of being spatially targeted, and the number of projects increases as poverty rates rise. Finally, the results of the subgroup analyses suggest that projects target multiple types of neighborhoods in terms of poverty, and this can potentially connect low-resourced neighborhoods with high-resourced neighborhoods via organizations. Based on this study, utilizing network analysis to examine organization and neighborhood networks can provide an alternative way to explore the nuances of CDBG project allocation and inform best practices in carrying out that process.