Towards Reduced Administrative Burdens: Performance Management, Machine Learning, and Evidence Building in the Federal Government
Loading...
Links to Files
Permanent Link
Collections
Author/Creator
Author/Creator ORCID
Date
2022-06-29
Type of Work
Department
University of Baltimore. College of Public Affairs
Program
University of Baltimore. Doctor of Public Administration.
Citation of Original Publication
Rights
Attribution-NoDerivs 3.0 United States
This item may be protected under Title 17 of the U.S. Copyright Law. It is made available by the University of Baltimore for non-commercial research and educational purposes.
This item may be protected under Title 17 of the U.S. Copyright Law. It is made available by the University of Baltimore for non-commercial research and educational purposes.
Abstract
Administrative burdens can inhibit how individuals access government services and benefits and reduce the effectiveness of programs. These burdens, including learning, compliance, and psychological costs, are experienced disparately and may cause greater negative impacts on individuals who would benefit most from the programs. This is especially true when they are “hidden” within policy and program design and implementation, thereby avoiding normal administrative law procedures that would allow citizens awareness and feedback opportunities. This study explores how existing standards for performance management and measurement through data analytics in the federal government can be leveraged to identify and measure administrative burdens with the aim of reducing their impact. This work focuses on the use of machine learning to solve implementation problems like administrative burdens, complete with the design, development, and implementation considerations that are specific to machine learning in the United States federal government. Finally, this work explores how to use the existing requirements for evidence-based policymaking and evaluation in the federal government to determine the impact of the machine learning solutions, as well as the impact of the reduced administrative burdens on the outcomes and goals of programs.
The approach of this research is to build on the existing academic literature, federal government requirements, and guidance to create three frameworks: identification and measurement, machine learning solutions, and evidence-building evaluations. Framework 1 provides a path to identifying, defining, and measuring administrative burdens within performance management processes. Framework 2 shows how to incorporate the nascent Federal government principles and guidance with academic and industry best practices to design, develop, and implement machine learning solutions in the public sector to reduce administrative burdens, which are identified and measured by Framework 1. Framework 3 is a guide to using existing federal government evidence-building and evaluation guidance to evaluate the implementations and impacts of the machine learning solutions and reduced administrative burdens.
This research demonstrates that administrative burdens in the federal government systems and processes can be identified and addressed without new legislation, regulation, or resources and that machine learning techniques are poised to provide solutions to public problems. This presents an opportunity for the federal government to refocus on providing performance data, administrative data, and information about existing uses of machine learning available to the public in a way that can benefit the academic research field; this lack of availability has follow-on impacts on the public sector. Additionally, this study shows that the field of administrative burden research needs to adopt shared definitions, measurement criteria, and approaches in order to build on existing theory and case studies and to magnify the impact of this research on the public sector. This research provides standard definitions of administrative burdens classification and measurement; a guide for agencies to reduce administrative burdens with performance management; practical guidance for applied machine learning in the federal government; an extension of evidence-based policy and evaluation research to focus on applied machine learning as well as the impact of reduced administrative burdens on program outcomes. These contributions benefit researchers focused on administrative burdens and provide practical support to government administrators.