Dr. Joshua Goldstein - UVa

Friday, October 22, 2021




This Colloquium will be held on line/Zoom, Friday, October 22nd - at 2:00PM

Measuring the Impacts of Rural Broadband Programs

Abstract: As a part of the American Recovery and Reinvestment Act of 2009, $2.9 billion was allocated to the USDA Rural Utilities Service (RUS) focused on broadband deployment in rural areas of the country. The Broadband Initiatives Program (BIP) program awarded grants, loans and grant/loan combinations to fund the development of broadband infrastructure.  To meet the BIP eligibility criteria, at least 75 percent of the funded area must have qualified as a rural area without sufficient access to broadband service. However, little has been done on evaluatingresearchers have not comprehensively evaluated the impacts of the BIP program and other USDA broadband funding programs.

My research uses information about BIP program awardees, property values, and socio-demographic factors by linking property sales data from CoreLogic, summaries from the American Community Survey (ACS) at the Census tract level, and FCC Form 477 data on broadband availability. I look at how these variables change across time in areas that have received RUS funding, areas that meet the program eligibility criteria but have not yet been funded, and ineligible area. I build on this analysis with the development of statistical models to examine the impact of the BIP using two methods of measuring change across program boundaries over time. The first is a spatial regression discontinuity approach, comparing businesses just inside the program area to businesses just outside the program boundary before and after the program was implemented. The second approach makes the same comparison using a spatial gradients method to assess change over time across the program boundary. These methods provide a novel evaluation of the BIP, and more broadly speaking, are general-purpose tools for program evaluation using micro-level data. I investigate not only the average impacts of the program, but impacts of specific BIP projects, by region, by the level of rurality across the rural-urban continuum and impacts of different broadband technologies promoted by BIP projects. Understanding the heterogeneity of such an extensive program produces valuable information for the design of current and future programs.

Bio: Josh Goldstein is a Research Assistant Professor at the Social and Decision Analytics Laboratory within the Biocomplexity Institute & Initiative of UVA. He received his PhD in statistics from Penn State in 2015. His research includes spatial and predictive modeling, Markov chain Monte Carlo methods, and generating synthetic populations. The population scale and dynamics of his research varies from the cellular level (respiratory syncytial virus infections) to country-wide populations in Africa (rotavirus infections in Africa) to county-level data in the U.S. (epidemics and invasive species).