Applying organizational density to local public service performance: separating homeless service outcomes from outputs

ABSTRACT

What happens to local services’ performance when service-provider density increases in a community? The answer is difficult. To explore how density relates to multiple aspects of performance, this study aims to examine the effects of service-providers’ density on service outputs and policy outcomes. Using a panel dataset of local homeless service planning bodies, the Continuum of Care Programme, we found that service outputs improved; however, the prevalence of homelessness did not decrease. Drawing upon organizational density theory, our findings contribute to the extant knowledge on public management by exploring how service-provider density relates to service outputs and policy outcomes separately.

KEYWORDS:

• Organizational density
• local homeless services performance
• two-way fixed effects model

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. Although there were 398 CoCs in 2018, 20 local CoCs that were smaller than a county were merged into one entity in order to use county-level control variables. This resulted in an analytical sample of 378 CoCs.

2. To illustrate this, consider two scenarios. In Scenario A, 365 people experience homelessness in a given year, but only for one night. In Scenario B, 365 people also experience homelessness in a given year, but for an entire year. In Scenario A, the average number of people experiencing homelessness on a given night would only be one whereas in Scenario B it would be 365.

3. What we refer to as service providers are called ‘organizations’ in the raw housing inventory count provided by the HUD.

4. While estimating regressions, we discovered one CoC, CA-523, to be very unlike other CoCs in terms of density and had an unbounded influence, a statistics term for one data point or a few data points changing the results of regressions in important ways. As such, we have dropped its observation from the data.

5. Available at http://ukcpr.org/resources/national-welfare-data, last accessed on 1 December 2019.

6. As a part of the present study, we also conducted a series of Granger causality tests to support this assumption, which indicated that lags in outcomes were jointly insignificant.

Additional information

Notes on contributors

Andrew Sullivan

Andrew Sullivan is an assistant professor of public administration at Southern Illinois University Edwardsville. His research focuses on competition and collaboration in federal-local partnerships, homelessness, and public finance.

Saerim Kim

Saerim Kim is an assistant professor of public administration and policy at Columbus State University. Her research involves understanding how cross-sectoral collaboration affects the performance of public service provision with special attention to financial resource composition and equity.

David Lee

David Lee is an assistant professor of public administration at the Ewha Womans University, Seoul, Republic of Korea. His research interests lie in collaborative public management and program evaluation.